Prenatal exposure to organophosphate esters and early neurodevelopment

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content PRENATAL EXPOSURE TO ORGANOPHOSPHATE ESTERS AND EARLY NEURODEVELOPMENT by Ixel C. Hernandez-Castro A Dissertation Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (EPIDEMIOLOGY) August 2023 Copyright 2023 Ixel C. Hernandez-Castro ii DEDICATION To my parents, Martha and Ricardo, for their immeasurable sacrifices and unwavering support. And to my husband, Ethan, for embracing every single one of my dreams, fervently believing in me, and pushing me through yet another finish line. iii ACKNOWLEDGEMENTS I am eternally grateful to the following people for their guidance, support, and encouragement throughout the process of completing this dissertation: To my dissertation chair, Dr. Theresa Bastain, for her invaluable mentorship, guidance, and encouragement throughout these past years. I am deeply indebted to you for your continual advocacy, feedback, and patience and for being my biggest champion. Thank you for investing so much time, energy, and resources into my professional development, and for always being willing to take a risk on me. To my dissertation committee members, Dr. Carrie Breton, Dr. Sandrah Eckel, Dr. Max Aung, and Dr. Brendan Grubbs, for your thoughtful feedback and expertise every step of the way. Dr. Carrie Breton, I am incredibly grateful for the expertise you continually shared and for encouraging me to think more critically about different perspectives and important biological pathways. Dr. Sandrah Eckel, thank you for always being so willing to provide critical statistical feedback and for sharing your incredible expertise in a wealth of statistical methodologies. Dr. Aung, I am extremely grateful for your unrelenting support, incredibly insightful feedback, and for always encouraging me to consider the implications of my work beyond the scientific world. Dr. Brendan Grubbs, thank you for providing your expertise on biological plausibility and for always being so encouraging throughout this process. To the amazing MADRES investigators whom I consider honorary committee members because of their substantial feedback, contributions, and support throughout this dissertation. Dr. Caitlin G. Howe, I cannot begin to express my thanks for all invaluable support, guidance, and expertise you provided on mixtures analysis throughout my dissertation and for continuously providing such thoughtful and constructive feedback on my papers. Dr. Rima Habre, words cannot express how thankful I am for the wealth of professional and personal wisdom you have shared, from impromptu SAS and JMP coding lessons, to sharing your incredible exposure science expertise, and for always cheering me on throughout this process. Dr. Shohreh Farzan and Dr. Claudia Toledo-Corral, thank you both for the many lessons you iv have shared, both professionally and personally, for always being so willing to share your incredible epidemiologic expertise, and for your continual encouragement. To my amazing PhD colleagues, whose encouragement and mentorship have guided me through many dissertation milestones. Dr. Alicia Peterson, thank you for taking me under your wing from the very beginning, holding my hand through all my dissertation related challenges, and being one of my biggest supporters. Dr. Karl O’Sharkey, I am eternally grateful for the invaluable support and instrumental wisdom you have shared with me throughout the entirety of this program and for selflessly helping me prepare for every dissertation milestone. I would also like to express my deepest gratitude to Dr. Christine Naya, Sahra Mohazzab-Hosseinian, Yan Xu, Yuhong Hu, Xiaoran Yang, and Noelle Pardo for their continual encouragement and support throughout this process. To the many wonderful staff and faculty in the Department of Population and Public Health Sciences who have supported me throughout the process, including Dr. Zhongzheng (Jason) Niu, Dr. Luis Maldonado, Steve Howland, Vivien Le, Milena Amadeus, Carmen Chavez, Rene Stanley, Celia Cedillo, Alejandra Valenzuela, Tingyu Yang, Marisela Miranda, Dr. Roberta McKean-Cowdin, Dr. Kimberly Siegmund, and Dr. Victoria Cortessis. A special thank you to the additional co-authors on this work for their constructive feedback, including Dr. Kurunthachalam Kannan, Morgan Robinson, Dr. Helen Foley, Dr. Laila Al-Marayati, Dr. Deborah Lerner, Dr. Nathana Lurvey, and Dr. Genevieve Dunton. Many thanks to the participants in the MADRES study, the MADRES investigator team, community partners, and the MADRES staff and data team without whom this project would not have been possible. And last, but certainly not least, thank you to my family and friends, Ethan Castro, Martha Hernandez, Ricardo Hernandez, Desiree Castro, Emmett Castro, Flor Hernandez, Richie Hernandez, Angel Hernandez, Elizabel Ortiz, Andres Ortiz, Jacqueline Gutierrez Garrett, Bibiana Martinez, Steven De La Torre, Carol Ochoa, Jose Gomez, Antonio Rosales, Eduardo Perez, and Edgar Lopez, for their undying support. v TABLE OF CONTENTS DEDICATION ...............................................................................................................................................ii ACKNOWLEDGEMENTS ......................................................................................................................... iii LIST OF TABLES ..................................................................................................................................... viii LIST OF FIGURES ...................................................................................................................................... xi ABSTRACT ................................................................................................................................................ xv INTRODUCTION ......................................................................................................................................... 1 REFERENCES .......................................................................................................................................... 6 CHAPTER 1 .................................................................................................................................................. 9 Early Neurodevelopment and Environmental Health Disparities ............................................................. 9 Organophosphate Esters (OPEs): Properties, Environmental Prescence, and Human Exposures .......... 11 Prenatal OPE Exposures and Direct Biomechanisms of Neurotoxicity .................................................. 16 Prenatal OPE Exposures and Indirect Biomechanisms of Neurotoxicity ............................................... 18 Birth Outcomes as Early Indicators of Later Neurodevelopmental Outcomes ....................................... 19 Prenatal OPE Exposures and Biomechanisms Contributing to Adverse Birth Outcomes ...................... 20 Epidemiological Evidence of Early Developmental Impacts .................................................................. 21 Neurodevelopment .............................................................................................................................. 21 Birth Outcomes ................................................................................................................................... 24 Current Gaps in the Literature on OPEs and Neurodevelopment ........................................................... 26 REFERENCES ........................................................................................................................................ 29 CHAPTER 2 ................................................................................................................................................ 40 Methods Overview .................................................................................................................................. 40 MADRES Cohort Description ............................................................................................................ 40 Relevant Methodology ............................................................................................................................ 42 Measures of Gestational Age at Birth and Birthweight-for-GA Z scores........................................... 44 Gross and Fine Motor Scores During Infancy .................................................................................... 44 Child Behavior Checklist (CBCL) ...................................................................................................... 45 MADRES Funding .................................................................................................................................. 46 REFERENCES ........................................................................................................................................ 47 CHAPTER 3 ................................................................................................................................................ 48 ABSTRACT ............................................................................................................................................ 48 INTRODUCTION ................................................................................................................................... 49 METHODS .............................................................................................................................................. 51 Study Sample ...................................................................................................................................... 51 OPE Metabolite Analysis .................................................................................................................... 52 Birth Outcomes ................................................................................................................................... 53 Covariates............................................................................................................................................ 53 Statistical Analysis .............................................................................................................................. 55 vi RESULTS ................................................................................................................................................ 58 Participant Characteristics................................................................................................................... 58 Individual Metabolite Associations with Birthweight and Gestational Age at Birth ......................... 59 Associations of OPE mixtures with BW for GA z-scores and Gestational Age at Birth ................... 61 DISCUSSION ......................................................................................................................................... 64 CONCLUSIONS ..................................................................................................................................... 70 SUPPLEMENTAL MATERIALS .......................................................................................................... 82 CHAPTER 4 .............................................................................................................................................. 123 ABSTRACT .......................................................................................................................................... 123 INTRODUCTION ................................................................................................................................. 124 METHODS ............................................................................................................................................ 125 Study Design ..................................................................................................................................... 125 OPE Metabolite Analysis .................................................................................................................. 126 Health Outcome Assessment ............................................................................................................ 127 Covariates.......................................................................................................................................... 128 Statistical Analysis ............................................................................................................................ 129 RESULTS .............................................................................................................................................. 131 Descriptive Statistics ......................................................................................................................... 131 Prenatal OPE Metabolite Concentrations ......................................................................................... 131 Motor Development Assessments ..................................................................................................... 131 Prenatal OPE Metabolites and Motor Development Across Infancy ............................................... 132 DISCUSSION ....................................................................................................................................... 134 CONCLUSION ..................................................................................................................................... 136 SUPPLEMENTAL MATERIALS ........................................................................................................ 145 REFERENCES ...................................................................................................................................... 154 CHAPTER 5 .............................................................................................................................................. 159 ABSTRACT .......................................................................................................................................... 159 INTRODUCTION ................................................................................................................................. 160 METHODS ............................................................................................................................................ 162 Study Design ..................................................................................................................................... 162 OPE Metabolites ............................................................................................................................... 163 Health Outcome Assessment ............................................................................................................ 164 Covariates.......................................................................................................................................... 165 Statistical Analysis ............................................................................................................................ 166 Sensitivity Analysis........................................................................................................................... 169 RESULTS .............................................................................................................................................. 169 Descriptive Statistics ......................................................................................................................... 169 Individual Metabolite Associations .................................................................................................. 170 Mixtures Associations ....................................................................................................................... 172 DISCUSSION ....................................................................................................................................... 174 CONCLUSION ..................................................................................................................................... 179 vii SUPPLEMENTAL FIGURES .............................................................................................................. 191 REFERENCES ...................................................................................................................................... 208 CHAPTER 6 .............................................................................................................................................. 213 SUMMARY .......................................................................................................................................... 213 IMPLICATIONS FOR POLICY AND PRACTICE............................................................................. 215 FLAME RETARDANTS AND NEUROTOXICITY: A CASE STUDY ON A REPETITIVE CYCLE OF REGRETTABLE SUBSTITUTIONS............................................................................................. 216 FUTURE STUDIES AND RESEARCH DIRECTIONS ...................................................................... 221 CONCLUSIONS ................................................................................................................................... 222 REFERENCES ...................................................................................................................................... 223 viii LIST OF TABLES Table 1.1. Organophosphate Ester (OPE) Parent Compounds and Corresponding Metabolites Analyzed…………………………………………………………….……….14 Table 1.2. OPEs Physiological Properties and Uses……………………………………………………....15 Table 1.3. OPEs Metabolites Physical and Physiological Properties………………………………….….16 Table 2.1. MADRES Cohort Participant Characteristics (N=774)…….………..…………………………39 Table 3.1. Participant Characteristics (N=421)………………………………………………………..…..77 Table 3.2. Distribution of Specific Gravity Adjusted OPE Concentrations (ng/mL) in Maternal Urine (N=421)………………….……………………………………………………………78 Table 3.3. Associations Between Individual OPE Urinary Metabolites (ng/mL) and Gestational Age at Birth (weeks).…………………………………………………………………………79 Table 3.4. Associations Between Individual OPE Urinary Metabolites (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores…………………………………...……………..80 Table 3.5. Posterior Inclusion Probabilities (PIPs) for OPE Urinary Metabolites and Gestational Age at Birth and Birthweight for Gestational Age (BW for GA) Z-scores Across Full and Sex-Stratified BKMR Mixture Models….………………………………………………81 Table S.3.1. Median Concentrations (ng/mL) of Urinary OPE Metabolites Across Published Studies………………………………………………………………….……………..103 Table S.3.2a. Prenatal OPE Urinary Metabolite Concentrations By Selected Sample Collection Variables (N=421)……………………………………………………………………………105 Table S.3.2b. Prenatal OPE Urinary Metabolite Concentrations By Selected Maternal Characteristics (N=421)………………………………………………………………………………….106 Table S.3.2c. Prenatal OPE Urinary Metabolites By Selected Infant Characteristics………………..….108 Table S.3.3. Associations between Prenatal OPE Urinary Metabolite Concentrations (ng/mL) and Gestational Age at Birth in Participants Who Reported No In-Utero Smoking (N=413)………………………………………………………………………………………..109 Table S.3.4. Associations between Prenatal OPE Urinary Metabolite Concentrations (ng/mL) and Gestational Age at Birth in Participants Who Reported No In-Utero Smoking (N=413)………………………………………………………………………………………..110 Table S.3.5. Associations Between Prenatal Urinary OPE Concentrations (ng/mL) and Gestational Age (GA) at Birth, Additionally Adjusting for Gestational Diabetes (N=421)…………………………………………………………………………………..………………111 Table S.3.6. Associations between Prenatal Urinary OPE Concentrations (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Additionally Adjusting ix for Gestational Diabetes (N=421)………………………………………………………………………..112 Table S.3.7. Associations Between Prenatal Urinary OPE Concentrations (ng/mL) and Gestational Age (GA) at Birth, Additionally Adjusting for Delivery Method (N=420)…………………………… ……………………………………………………………….……113 Table S.3.8. Associations between Prenatal Urinary OPE Concentrations (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Additionally Adjusting for Delivery Method (N=420)……………………………….. ………………………………………….114 Table S.3.9. Single Exposure Effect Estimates for Each Prenatal OPE Metabolite in the Bayesian Kernel Machine Regression (BKMR) Mixture and Gestational Age (GA at Birth) Model……………………………………………………………………………………...115 Table S.3.10. Single Exposure Effect Estimates for Each Prenatal OPE Metabolite in the Bayesian Kernel Machine Regression (BKMR) Mixture and Birthweight for Gestational Age (BW for GA) Z-score Model………………………………...…………………………116 Table 4.1. Participant Characteristics (N=329)………………………………………………………..…140 Table 4.2. Distribution of Specific Gravity Adjusted OPEs Concentrations (ng/mL) in Maternal Urine (N=329)…………...………………………………………………………………….141 Table 4.3. Longitudinal Associations Between Prenatal OPEs and ASQ-3 Motor Development Across Infancy (N=329)……………………...………...…………………………………142 Table 4.4. Longitudinal Associations Between Prenatal OPEs and ASQ-3 Motor Development Groups (“High Risk of Delay” vs “Typically developing”) Across Infancy, by Infant Sex………………………………………………………………………………..…..143 Table 4.5. Longitudinal Associations Between Prenatal OPEs and ASQ-3 Motor Development Counts, by Infant Sex………….………………………………………………………….144 Table S.4.1. Comparison of Participant Characteristics in the Full Analytical Sample vs those Excluded from the Sample Due to Missing Information on Key Variables…………………………………………………………………………………………...…….148 Table S.4.2. Cross-Sectional Associations Between Prenatal OPEs and ASQ-3 Motor Development Groups (“Typical” vs “High Risk of Delay”) Across Infancy……………………………………………………………………………………….…………..149 Table S.4.3. Cross-Sectional Associations Between Prenatal OPEs and ASQ-3 Motor Counts Across Infancy…………………………………….……………………………………...150 Table S.4.4. Longitudinal Associations Between Prenatal OPEs and ASQ-3 Motor Development Across Infancy Excluding Participants that Smoked In- Utero (N=329)……………………………………..…………………………………………………..…151 Table S.4.5. Longitudinal Associations Between Prenatal OPEs and ASQ-3 Motor Development Across Infancy Among Full Term Births Only (≥37 weeks GA at Birth) (N=301)……………………..………………………………………………...152 x Table S.4.6. Longitudinal Associations Between Prenatal OPEs and ASQ-3 Motor Development Across Infancy, When Rounding Non-Integers Up vs Down (N=329)…………………………………… …………………………………………………………….153 Table 5.1. Participant Characteristics (N=204)…………………………………………………………..186 Table 5.2. Distribution of Specific Gravity Adjusted OPE Concentrations (ng/mL) in Maternal Urine (N=204)………………………………………….………………………….187 Table 5.3. Individual Associations Between Third Trimester Urinary OPE Metabolites (ng/mL) and CBCL Raw Composite Scores (N=204)……. ……………………………….188 Table 5.4. Posterior Inclusion Probabilities (PIPs) and Single Exposure Effect Estimates for Each Prenatal OPE Metabolite in the Bayesian Kernel Machine Regression (BKMR) Mixture and CBCL Composite Raw Score.............……………………………….190 Table S.5.1. Distribution of Specific Gravity Adjusted OPE Concentrations (ng/mL) in Urine for Maternal Participants Analyzed (N=204) vs the Full Sample of Maternal Participants with OPEs Available (N=426)………….…………………………….202 Table S.5.2. Comparison of Participant Characteristics Analyzed in the Analytical Dataset (N=204) to Subset with OPE Metabolite Concentrations Available (N=426) and Full MADRES Participants who Have Delivered Children in the Study (N=774)……………………………………………………………………...……203 Table S.5.3. Individual Associations Between Third Trimester Urinary OPE Metabolites (ng/mL) and CBCL Composite T-Scores (n=204) ......……………………………………..205 Table S.5.4. Individual Associations Between Third Trimester Urinary OPE Metabolites (ng/mL) and CBCL Raw Composite Scores Among Participants Who Reported No In-Utero Smoking (N= 199).………………………………………………………………...…………………….207 Table 6.1. Prior Phase-Out Challenges and Recommendations of Potential Solutions Moving Forward………………………………………………………………………………………………..…220 xi LIST OF FIGURES Figure 1.1. OPE Exposures, Mechanisms, and Existing Observational Evidence……………………….28 Figure 3.1. Spearman Correlations of Organophosphate Ester Metabolites (ng/mL) in Third Trimester Maternal Urine ………..…………………………………………………………………71 Figure 3.2. Associations Between Prenatal OPE Urinary Metabolite Concentrations (ng/mL) and Gestational Age at Birth, Using Generalized Additive Models (GAMs)…...………………72 Figure 3.3. Associations Between Prenatal OPE Metabolite Concentrations (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Using Generalized Additive Models (GAMs)……………….………………………………………………………………………..…73 Figure 3.4. OPE Metabolite Mixtures (ng/mL) and Infant Gestational Age at Birth (weeks) and Birthweight for Gestational Age (BW for GA) Z Scores in Full Models, Using BKMR (N=421)……….74 Figure 3.5. OPE Metabolite Mixtures (ng/mL) and Infant Gestational Age at Birth (weeks) and Birthweight for Gestational Age (BW for GA) Z Scores in Female Stratified Models, Using BKMR (N=215)…. ……………………………………………………………………………………….75 Figure 3.6. OPE Metabolite Mixtures (ng/mL) and Infant Gestational Age at Birth (weeks) and Birthweight for Gestational Age (BW for GA) Z Scores in Male Stratified Models, Using BKMR (N=206)…………………………………………………………………………………………...……….76 Figure S.3.1a. Directed Acyclic Graph (DAG) for Associations Between Prenatal OPE Urinary Metabolites and Gestational Age (GA) at Birth……………………...……………………………………82 Figure S.3.1b. Directed Acyclic Graph (DAG) for Associations Between Prenatal OPE Urinary Metabolites and Birthweight………………………………………………………………….…..83 Figure S.3.2. Associations Between Prenatal OPE Urinary Metabolite Concentrations (ng/mL) and Gestational Age (GA) at Birth (weeks), Using Generalized Additive Models (GAMs) (Non-smoking Participants Only (n=413))….…………………………………………………84 Figure S.3.3. Associations Between Prenatal OPE Metabolite Concentrations (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Using Generalized Additive Models (GAMs) (Non-smoking Participants Only (n=413)) ……………………………………………..85 Figure S.3.4. Associations Between Prenatal OPE Metabolite Concentrations (ng/mL) and GA at Birth, Using Generalized Additive Models (GAMs) (Additional Adjustment for Gestational Diabetes Mellitus (n=421)) ………………………………………………………………...…………......86 Figure S.3.5. Associations Between Prenatal OPE Metabolite Concentrations (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Using Generalized Additive Models (GAMs) (Additional Adjustment for Gestational Diabetes Mellitus (n=421)) ………………...…………87 Figure S.3.6. Associations Between Prenatal OPE Metabolite Concentrations (ng/mL) and GA at Birth, Using Generalized Additive Models (GAMs) (Additional Adjustment for Delivery Method (n=420))………..…………………………………………………………………………………88 xii Figure S.3.7. Associations Between Prenatal OPE Metabolite Concentrations (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Using Generalized Additive Models (GAMs) (Additional Adjustment for Delivery Method (n=420)) ………….…………………….89 Figure S.3.8. Bivariate Associations Between OPE Urinary Metabolite Mixtures (ng/mL) and Gestational Age (GA) at Birth and Birthweight for Gestational Age (BW for GA) Z-Scores, Using Bayesian Kernel Machine Regression (BKMR) ………………………………………………………….90 Figure S.3.9. Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Infant Gestational Age (GA) at Birth (weeks) in Full and Sex-Stratified Models, Using Bayesian Kernel Machine Regression (BKMR) (Non-smoking Participants Only (n=413)) …………………………………...……91 Figure S.3.10. Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores in Full and Sex-Stratified Models, Using Bayesian Kernel Machine Regression (BKMR) (Non-smoking Participants Only (n=413)) ……………..………..92 Figure S.3.11. Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Gestational Age (GA) at Birth (weeks) Varying the Smoothing Parameter to b=50 and b=1000, Using Bayesian Kernel Machine Regression (BKMR) (Full Model N=421) ………………………………………………..……..93 Figure S.3.12. Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Gestational Age (GA) at Birth (weeks) Varying the Smoothing Parameter to b=50 and b=1000, Using Bayesian Kernel Machine Regression (BKMR) (Female Stratified N=215) …………………………………...……………94 Figure S.3.13. Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Gestational Age (GA) at Birth (weeks) Varying the Smoothing Parameter to b=50 and b=1000, Using Bayesian Kernel Machine Regression (BKMR) (Male Stratified N=206) ………….……….………………………………95 Figure S.3.14. Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Using Bayesian Kernel Machine Regression (BKMR) Varying the Smoothing Parameter to b=50 and b=1000 (Full Model N=421) …………………………….96 Figure S.3.15. Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Using Bayesian Kernel Machine Regression (BKMR) Varying the Smoothing Parameter to b=50 and b=1000 (Female Stratified N=215) ………………………97 Figure S.3.16. Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Using Bayesian Kernel Machine Regression (BKMR), Varying the Smoothing Parameter to b=50 and b=1000 (Male Stratified N=206) …………….…………..98 Figure S.3.17. Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Infant Gestational Age (GA) at Birth (weeks), Using Bayesian Kernel Machine Regression (BKMR), in Full and Sex-Stratified Models (Additionally Adjusting for Gestational Diabetes Mellitus (n=421)).…………99 Figure S.3.18. Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Using Bayesian Kernel Machine Regression (BKMR) in Full and Sex-Stratified Models (Additionally Adjusting for Gestational Diabetes Mellitus (n=421))……………………………………………………………………….…………………………100 Figure S.3.19. Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Infant Gestational Age (weeks), Using Bayesian Kernel Machine Regression (BKMR) in Full and Sex-Stratified xiii Models (Additionally Adjusting for Method of Delivery (n=421)) ……………………………….……...101 Figure S.3.20. Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Using Bayesian Kernel Machine Regression (BKMR) in Full and Sex-Stratified Models (Additionally Adjusting for Method of Delivery (n=421)) …………...102 Figure 4.1. Consort Diagram of Included Mother-Infant Dyads…………………………………………137 Figure 4.2. Spearman Correlations of Organophosphate Ester Metabolite Concentrations in Third Trimester Maternal Urine. ………...…………………………………………………………….138 Figure 4.3. Distribution of Ages and Stages (ASQ-3) Gross and Fine Motor Subscale Scores by Study Timepoint (N=329) ………………………………………………………………….…139 Figure S.4.1. Directed Acyclic Graph (DAG) of Prenatal OPE Metabolites and Infant Motor Development……………………………………………………………………………..………………145 Figure S.4.2. Associations Between Prenatal OPE Metabolite Concentrations (ng/mL) and Gross and Fine Motor ASQ-3 Groups…………………………………….……………………………………..146 Figure S.4.3. Associations Between Prenatal OPE Metabolite Concentrations (ng/mL) and Gross and Fine Motor ASQ-3 Counts………...…………………………………………………………………147 Figure 5.1. Consort Diagram of Included Mother-Infant Dyads……………………………………..…..180 Figure 5.2. Spearman Correlations of Organophosphate Ester Metabolites (ng/mL) in Third Trimester Maternal Urine………………………………………………………………………………...181 Figure 5.3. Distributions of 36 Month Child Behavior Checklist Composite Raw Scores (CBCL) (N=204) ………………………………………………………………………………………...182 Figure 5.4. Associations Between Urinary Prenatal OPE Metabolite Concentrations (ng/mL) and CBCL Composite Raw Scores, Using Generalized Additive Models (N=204) ……...………………183 Figure 5.5. Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and CBCL Composite Raw Scores, Using BKMR (n=204)……....……………………………………………………………………184 Figure 5.6. Bivariate Associations Between Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and CBCL Composite Raw Scores, Using BKMR (n=204)...………………………………………….…185 Figure S.5.1. Directed Acyclic Graph (DAG) of Prenatal OPE metabolites and Child Neurobehavioral Development..………….………………………………………………………………191 Figure S.5.2. Associations Between Urinary Prenatal OPE Metabolite Concentrations (ng/mL) and CBCL Composite T-Scores, Using Generalized Additive Models (N=204) ……………………...…193 Figure S.5.3. Associations Between Urinary Prenatal OPE Metabolite Concentrations (ng/mL) and CBCL Composite Raw Scores Among Participants Who Reported No In-Utero Smoking, Using Generalized Additive Models (n=199) ………………………………………………………...….194 Figure S.5.4. Posterior Inclusion Probabilities (PIPs) for Pairwise Interactions Between OPE xiv Metabolites and CBCL Composite Raw Scores Using NLinteraction Method…………...………………195 Figure S.5.5. Prenatal DNBP+DIBP Exposures and Children’s Total Problems Scores by Tertiles of BCEP, Using Generalized Additive Models…………………………………………………….……..196 Figure S.5.6. Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and CBCL Composite T-Scores, Using BKMR (N=204) …………………………………………….….…………………………………197 Figure S.5.7. Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and CBCL Composite Raw Scores Among Participants Who Reported No In-Utero Smoking, Using BKMR (N=199) …….….198 Figure S.5.8. Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and CBCL Composite Raw Scores, Using BKMR Varying the Smoothing Parameter to b=50………………………………….199 Figure S.5.9. Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and CBCL Composite Raw Scores, Using BKMR Varying the Smoothing Parameter to b=1000…………………………...…..200 Figure S.5.10. Posterior Inclusion Probabilities (PIPs) for Pairwise Interactions Between OPE Metabolites and CBCL Composite Raw Scores Using NLinteraction Method and Increasing the Threshold to 0.25…………………………………………………………………………………………201 Figure 6.1. Challenges Experienced During Prior Phase Outs Which Contributed to Use of PBDEs and OPEs…..……………………………………………………………………………………..219 xv ABSTRACT Early neurodevelopment begins in-utero with the formation of integral neurological structures and mechanisms and plays a critical role in supporting children’s lifelong health. According to previous studies, ubiquitous environmental chemicals, such as organophosphate esters (OPEs) used as flame retardants and plasticizers in a variety of consumer products, may pose a neurotoxic risk at environmentally relevant doses. However, there is limited observational evidence evaluating the association between prenatal exposures to OPEs and early neurodevelopment, particularly among populations historically underrepresented in the biomedical sciences. Additionally, there is scarce evidence evaluating the impacts of co-occurring OPEs on early neurodevelopment, along with many OPEs whose impacts on neurodevelopment are not well understood. We investigated the impacts of prenatal OPE exposures on early neurodevelopmental outcomes among a predominately low-income and Hispanic pregnancy cohort of participants residing in Los Angeles, California, the Maternal and Developmental Risks from Environmental and Social Stressors (MADRES) cohort. We evaluated the association between nine OPE metabolites measured in urine samples collected during a third trimester visit and outcomes associated with neurodevelopmental health, including gestational age (GA) at birth, birthweight, motor development in infancy, and neurobehavioral symptoms in early childhood. Altogether, we found evidence of potentially neurotoxic impacts from prenatal OPE exposures on development, with various sex-specific associations, and interesting mixtures associations, which highlight the value of evaluating co-occurring OPE exposures on children’s health outcomes. These findings can potentially inform future reduction efforts to minimize adverse effects of prenatal OPEs on neurodevelopmental health but would benefit from further research on individual sources of OPEs along with more research using multiple measurements of OPEs across prenatal and postnatal windows. 1 INTRODUCTION Early behavioral and neurological development play critical roles in promoting healthy child development and lifelong health. 1 Lower performance on behavioral and neurodevelopmental assessments in childhood has been tied to increased cardiometabolic conditions and psychiatric disorders in adulthood along with increased violence, substance abuse, and antisocial behaviors. 2-10 Early delays in infant behavioral and neurological development may be indicative of later developmental disabilities which are a group of lifelong conditions that impair physical, cognitive, or behavioral functioning. 11,12 Developmental disabilities are common among US children, with an estimated cumulative prevalence of 16.9% from 2009- 2017. 12 The prevalence of developmental disabilities has increased over previous years, particularly among boys, children in non-Hispanic white and Hispanic families, and children living in urban areas with mothers with less than a college education, with an increase in cumulative prevalence of 9.5% from 2009-2011 to 2015-2017. 12 Subclinical deficits in early neurodevelopment are even more common than developmental disabilities, with adverse impacts on well-being, social functioning, academic achievement, and physical and mental health. 13 This is of significant public health concern since there are exorbitant social, economic, and health costs of developmental disabilities, with substantially higher estimated costs when considering potential subclinical deficits. 14-16 Thus, it is imperative to identify modifiable risks to early neurodevelopment, such as environmental chemical exposures, in order to inform appropriate prevention methods. Accumulating evidence suggests that environmental chemicals may impact the fetal environment and result in permanent changes in the anatomy and physiology of the offspring, also known as fetal programming, a critical component of the Developmental Origins of Health and Disease (DOHaD) hypothesis. 17 Previous literature suggests that prenatal exposures to environmental chemicals , particularly endocrine disrupting chemicals, can adversely impact early neurodevelopmental outcomes. 18-20 The developing central nervous system is particularly susceptible to environmental chemical exposures since 2 the structural foundation and formation of the brain begins and rapidly progresses in-utero, with a cascade of neurological processes resulting in multiple periods of vulnerability and susceptibility. 21,22 Neurodevelopmental processes during fetal and early life occur within a tightly controlled, sequenced framework, with incremental changes in development resulting from toxicant exposures having the potential to result in lifelong health effects. 23 However, only a small number of industrial chemicals are well-established neurotoxicants in children (~12), with many potential neurotoxic chemicals likely to lie undiscovered. 13 Among the small number of recognized neurotoxic chemicals in children, flame retardants, specifically polybrominated diphenyl ethers (PBDEs), have been found to adversely impact early neurodevelopment. 24-26 Studies further suggest higher exposures of PBDEs among low-income and Mexican-American or Mexican immigrant households, possibly stemming from socioeconomic barriers to acquiring PBDE free products, highlighting the disproportionate burden of flame retardant exposures experienced by structurally marginalized populations. 26,27 Fortunately, PBDEs have been phased out of use; however, replacement chemical flame retardants, such as organophosphate esters (OPEs), have rapidly increased in use. 28 OPEs are also commonly used as plasticizers and lubricants, contributing to their environmental ubiquity. 29,30 Although there is very limited epidemiological literature on the health effects of OPEs, emerging evidence suggests adverse neurodevelopmental impacts of prenatal OPE exposures among children. 31-34 Given the widespread exposure of OPEs among pregnant women, there is a critical need for epidemiological research to better understand the impacts of prenatal OPE exposures on neurodevelopmental outcomes. 35,36 Therefore, the primary aim of this dissertation is to evaluate the impacts of prenatal OPE metabolite concentrations on early neurodevelopment among a predominantly low-income Hispanic pregnancy cohort of mother-child pairs in Los Angeles. Given the synergistic and highly intertwined relationship between neurodevelopment and neurobehavior during early development, 37 the term “neurodevelopment” will be used throughout this dissertation to encompass both neurodevelopmental and neurobehavioral outcomes. However, the term “neurobehavior” will exclusively be used in research study 3 which specifically 3 discusses the neurobehavioral outcomes of that study. This dissertation will assess the effects of prenatal OPE metabolite exposures on key outcomes important for childhood neurodevelopment including birth outcomes, 38,39 motor development in infancy, 40 and emotional and behavioral health in early childhood. 41 Chapter 1 of this dissertation begins with an introduction on the importance of early neurodevelopmental outcomes and potential adverse effects of environmental exposures, which may disproportionately impact health disparities populations, especially during particularly susceptible periods of development. Following this, chemical flame retardants are briefly introduced, with a short outline on the history of chemical flame retardants and children’s neurodevelopment and a subsequent introduction of OPEs and their physiological properties, environmental presence, and human exposures. Hypothesized biological mechanisms of prenatal OPE exposures on direct and indirect neurotoxicity are then summarized, followed by a review of the literature on OPEs and early developmental impacts. In Chapter 2, relevant methods across all studies of this dissertation are introduced, including information on exposure measurement, demographics, and outcome measures. The study population analyzed in this dissertation, participants from the Maternal And Developmental Risks from Environmental and Social Stressors (MADRES) cohort, is also briefly introduced, along with information on the cohort’s participant characteristics and other critical study design information. In this dissertation’s first research paper (Chapter 3), which was published in Environmental Research, 42 we examined the association between prenatal OPE metabolites independently and as a mixture on gestational age (GA) at birth and birthweight for gestational age (BW-for-GA). We additionally assessed potential effect modification by infant sex, given prior evidence indicating potential differences in toxicity by sex. This study additionally characterized OPE exposures across demographic characteristics of the MADRES cohort. We hypothesized that both independent OPE metabolites and OPE metabolite mixtures would be associated with earlier GA at birth and lower BW-for-GA, with more adverse impacts on female infants. Using urinary OPE metabolites measured during a third trimester visit and birth outcomes primarily abstracted from electronic medical records, we performed multiple linear regression and generalized additive models (GAMs) to evaluate the individual impacts of prenatal OPEs on GA at birth and BW-for- 4 GA z-scores. We then performed Bayesian kernel machine regression (BKMR) to evaluate the associations between five OPE metabolites in mixtures and infant birth outcomes. This study found evidence of sex- specific impacts of OPEs on GA at birth, with BDCIPP associated with significantly earlier GA at birth among male infants only and DNBP+DIBP associated with earlier GA at birth at high concentrations among female infants only. In mixtures models, the cumulative OPE mixture was associated with earlier GA at birth at higher concentrations, with more pronounced associations observed among female infants. In the second research paper (Chapter 4), we further explored the neurodevelopmental impacts of prenatal OPEs in infancy by evaluating the association between prenatal OPE metabolite concentrations and gross and fine motor development at 6, 9, 12, and 18-months. Effect modification by infant sex was similarly assessed by using a statistical interaction term and stratifying models by infant sex. Mixed effects logistic regressions and negative binomial mixed models were performed to evaluate the risk of infant motor delay along with change in motor performance across time. This study found an association between prenatal DPHP concentrations and risk of potential fine motor delays and decreased fine motor performance. A potentially protective association between BMPP and gross motor development was also found. There was also evidence of sex-specific effects with DRPR and BCIPP and gross and fine motor development, with adverse patterns among female infants. The third research paper of this dissertation (Chapter 5), explored the association between prenatal OPEs and neurobehavioral symptoms at 36 months, including internalizing, externalizing, and total problems, along with effect modification by infant sex. We performed both single metabolite associations, using linear regression models and GAMs, along with mixtures models, using BKMR. This study found an association between detectable BMPP levels and higher externalizing and total problems in adjusted linear regression models. Sex-specific associations were also observed between BCIPP and the internalizing problems and total problems, with adverse associations among males only. Mixtures models were further suggestive of possible pairwise interactions between metabolites. Finally, in Chapter 6, a summary of this dissertation’s main findings and their implications for children’s health are discussed. We present a case study on flame retardant chemicals’ repetitive cycle of 5 regrettable substitutions and propose potential policy recommendations and alternative solutions. This chapter concludes with future research directions. 6 REFERENCES 1. Child NSCotD. Connecting the Brain to the Rest of the Body: Early Childhood Development and Lifelong Health are Deeply Intertwined. 2020. www.developingchild.harvard.edu: Harvard University. 2. Camargos ACR, Mendonça VA, Andrade CAd, et al. Overweight and obese infants present lower cognitive and motor development scores than normal-weight peers. Research in Developmental Disabilities 2016; 59: 410-416. DOI: https://doi.org/10.1016/j.ridd.2016.10.001. 3. Shoaibi A, Neelon B, Østbye T, et al. Longitudinal associations of gross motor development, motor milestone achievement and weight-for-length z score in a racially diverse cohort of US infants. BMJ Open 2019; 9: e024440. DOI: 10.1136/bmjopen-2018-024440. 4. Flensborg-Madsen T, Sørensen HJ, Revsbech R, et al. Early motor developmental milestones and level of neuroticism in young adulthood: a 23-year follow-up study of the Copenhagen Perinatal Cohort. Psychological Medicine 2013; 43: 1293-1301. 2012/09/14. DOI: 10.1017/S0033291712001997. 5. Manzardo AM, Penick EC, Knop J, et al. Developmental Differences in Childhood Motor Coordination Predict Adult Alcohol Dependence: Proposed Role for the Cerebellum in Alcoholism. Alcoholism: Clinical and Experimental Research 2005; 29: 353-357. https://doi.org/10.1097/01.ALC.0000156126.22194.E0. DOI: https://doi.org/10.1097/01.ALC.0000156126.22194.E0. 6. Schiffman J, Sorensen HJ, Maeda J, et al. Childhood Motor Coordination and Adult Schizophrenia Spectrum Disorders. American Journal of Psychiatry 2009; 166: 1041-1047. DOI: 10.1176/appi.ajp.2009.08091400. 7. Paradis AD, Fitzmaurice GM, Koenen KC, et al. A prospective investigation of neurodevelopmental risk factors for adult antisocial behavior combining official arrest records and self- reports. J Psychiatr Res 2015; 68: 363-370. 2015/06/08. DOI: 10.1016/j.jpsychires.2015.04.030. 8. Raine A. A neurodevelopmental perspective on male violence. Infant Ment Health J 2019; 40: 84-97. 2018/12/27. DOI: 10.1002/imhj.21761. 9. Gao Y, Glenn AL, Schug RA, et al. The neurobiology of psychopathy: A neurodevelopmental perspective. The Canadian Psychiatric Association Journal / La Revue de l'Association des psychiatres du Canada 2009; 54: 813-823. 10. Holland L, Reid N and Smirnov A. Neurodevelopmental disorders in youth justice: a systematic review of screening, assessment and interventions. Journal of Experimental Criminology 2021. DOI: 10.1007/s11292-021-09475-w. 11. Vlasblom E, Boere-Boonekamp MM, Hafkamp-de Groen E, et al. Predictive validity of developmental milestones for detecting limited intellectual functioning. PLoS One 2019; 14: e0214475. 2019/03/29. DOI: 10.1371/journal.pone.0214475. 12. Zablotsky B, Black LI, Maenner MJ, et al. Prevalence and Trends of Developmental Disabilities among Children in the United States: 2009–2017. Pediatrics 2019; 144: e20190811. DOI: 10.1542/peds.2019-0811. 13. Grandjean P and Landrigan PJ. Neurobehavioural effects of developmental toxicity. The Lancet Neurology 2014; 13: 330-338. DOI: 10.1016/S1474-4422(13)70278-3. 14. Trasande L and Liu Y. Reducing The Staggering Costs Of Environmental Disease In Children, Estimated At $76.6 Billion In 2008. Health Affairs 2011; 30: 863-870. DOI: 10.1377/hlthaff.2010.1239. 15. Polanczyk GV, Salum GA, Sugaya LS, et al. Annual research review: A meta‐analysis of the worldwide prevalence of mental disorders in children and adolescents. Journal of child psychology and psychiatry 2015; 56: 345-365. 16. Gustavsson A, Svensson M, Jacobi F, et al. Cost of disorders of the brain in Europe 2010. European neuropsychopharmacology 2011; 21: 718-779. 17. Schlotz W and Phillips DIW. Fetal origins of mental health: Evidence and mechanisms. Brain, Behavior, and Immunity 2009; 23: 905-916. DOI: https://doi.org/10.1016/j.bbi.2009.02.001. 7 18. Dórea JG. Environmental exposure to low-level lead (Pb) co-occurring with other neurotoxicants in early life and neurodevelopment of children. Environmental Research 2019; 177: 108641. DOI: https://doi.org/10.1016/j.envres.2019.108641. 19. Braun JM. Early-life exposure to EDCs: role in childhood obesity and neurodevelopment. Nat Rev Endocrinol 2017; 13: 161-173. 2016/11/20. DOI: 10.1038/nrendo.2016.186. 20. Jurewicz J, Polańska K and Hanke W. Chemical exposure early in life and the neurodevelopment of children--an overview of current epidemiological evidence. Ann Agric Environ Med 2013; 20: 465- 486. 2013/09/28. 21. Rauh VA and Margolis AE. Research Review: Environmental exposures, neurodevelopment, and child mental health - new paradigms for the study of brain and behavioral effects. J Child Psychol Psychiatry 2016; 57: 775-793. 2016/03/14. DOI: 10.1111/jcpp.12537. 22. Monk C and Hane A. Fetal and Infant Brain–Behavior Development: Milestones & Environmental Influences. 2014. 23. Stiles J and Jernigan TL. The basics of brain development. Neuropsychol Rev 2010; 20: 327-348. 2010/11/03. DOI: 10.1007/s11065-010-9148-4. 24. Costa LG and Giordano G. Developmental neurotoxicity of polybrominated diphenyl ether (PBDE) flame retardants. Neurotoxicology 2007; 28: 1047-1067. 2007/08/24. DOI: 10.1016/j.neuro.2007.08.007. 25. Vuong AM, Yolton K, Cecil KM, et al. Flame retardants and neurodevelopment: An updated review of epidemiological literature. Curr Epidemiol Rep 2020; 7: 220-236. 2021/01/08. DOI: 10.1007/s40471-020-00256-z. 26. Darrow LA, Jacobson MH, Preston EV, et al. Predictors of Serum Polybrominated Diphenyl Ether (PBDE) Concentrations among Children Aged 1–5 Years. Environmental Science & Technology 2017; 51: 645-654. DOI: 10.1021/acs.est.6b04696. 27. Eskenazi B, Fenster L, Castorina R, et al. A comparison of PBDE serum concentrations in Mexican and Mexican-American children living in California. Environ Health Perspect 2011; 119: 1442- 1448. 2011/04/15. DOI: 10.1289/ehp.1002874. 28. Blum A, Behl M, Birnbaum L, et al. Organophosphate Ester Flame Retardants: Are They a Regrettable Substitution for Polybrominated Diphenyl Ethers? Environ Sci Technol Lett 2019; 6: 638- 649. 2020/06/05. DOI: 10.1021/acs.estlett.9b00582. 29. Yang J, Zhao Y, Li M, et al. A Review of a Class of Emerging Contaminants: The Classification, Distribution, Intensity of Consumption, Synthesis Routes, Environmental Effects and Expectation of Pollution Abatement to Organophosphate Flame Retardants (OPFRs). Int J Mol Sci 2019; 20 2019/06/20. DOI: 10.3390/ijms20122874. 30. Hou R, Xu Y and Wang Z. Review of OPFRs in animals and humans: Absorption, bioaccumulation, metabolism, and internal exposure research. Chemosphere 2016; 153: 78-90. DOI: https://doi.org/10.1016/j.chemosphere.2016.03.003. 31. Castorina R, Bradman A, Stapleton HM, et al. Current-use flame retardants: Maternal exposure and neurodevelopment in children of the CHAMACOS cohort. Chemosphere 2017; 189: 574-580. 2017/10/01. DOI: 10.1016/j.chemosphere.2017.09.037. 32. Doherty BT, Hoffman K, Keil AP, et al. Prenatal exposure to organophosphate esters and cognitive development in young children in the Pregnancy, Infection, and Nutrition Study. Environ Res 2019; 169: 33-40. 2018/11/09. DOI: 10.1016/j.envres.2018.10.033. 33. Doherty BT, Hoffman K, Keil AP, et al. Prenatal exposure to organophosphate esters and behavioral development in young children in the Pregnancy, Infection, and Nutrition Study. Neurotoxicology 2019; 73: 150-160. 2019/04/06. DOI: 10.1016/j.neuro.2019.03.007. 34. Lipscomb ST, McClelland MM, MacDonald M, et al. Cross-sectional study of social behaviors in preschool children and exposure to flame retardants. Environ Health 2017; 16: 23. 2017/03/10. DOI: 10.1186/s12940-017-0224-6. 35. Hoffman K, Lorenzo A, Butt CM, et al. Predictors of urinary flame retardant concentration among pregnant women. Environ Int 2017; 98: 96-101. 2016/10/13. DOI: 10.1016/j.envint.2016.10.007. 8 36. Ospina M, Jayatilaka NK, Wong LY, et al. Exposure to organophosphate flame retardant chemicals in the U.S. general population: Data from the 2013-2014 National Health and Nutrition Examination Survey. Environ Int 2018; 110: 32-41. 2017/11/06. DOI: 10.1016/j.envint.2017.10.001. 37. Patel DR and Merrick J. Neurodevelopmental and neurobehavioral disorders. Transl Pediatr 2020; 9: S1-S2. DOI: 10.21037/tp.2020.02.03. 38. Luu TM, Rehman Mian MO and Nuyt AM. Long-Term Impact of Preterm Birth: Neurodevelopmental and Physical Health Outcomes. Clin Perinatol 2017; 44: 305-314. 2017/05/10. DOI: 10.1016/j.clp.2017.01.003. 39. You J, Shamsi BH, Hao M-C, et al. A study on the neurodevelopment outcomes of late preterm infants. BMC Neurol 2019; 19: 108-108. DOI: 10.1186/s12883-019-1336-0. 40. Heineman KR, Schendelaar P, Van den Heuvel ER, et al. Motor development in infancy is related to cognitive function at 4 years of age. Dev Med Child Neurol 2018; 60: 1149-1155. 2018/04/11. DOI: 10.1111/dmcn.13761. 41. Dubovický M. Neurobehavioral manifestations of developmental impairment of the brain. Interdiscip Toxicol 2010; 3: 59-67. DOI: 10.2478/v10102-010-0012-4. 42. Hernandez-Castro I, Eckel SP, Howe CG, et al. Sex-specific effects of prenatal organophosphate ester (OPE) metabolite mixtures and adverse infant birth outcomes in the maternal and developmental risks from environmental and social stressors (MADRES) pregnancy cohort. Environmental Research 2023; 226: 115703. DOI: https://doi.org/10.1016/j.envres.2023.115703. 9 CHAPTER 1 Early Neurodevelopment and Environmental Health Disparities About 1 in 6 children in the United States have been diagnosed with a developmental disorder, such as learning disabilities, sensory deficits, developmental delays, cerebral palsy, autism, and attention deficit/hyperactivity disorder (ADHD). 1 Over previous decades, the prevalence of developmental disorders has been increasing among children, particularly among boys, children in non-Hispanic white and Hispanic families, children with private insurance, and children living in urban areas with mothers with less than a college education. 1,2 The social and economic costs of such developmental disabilities are high and extensive efforts are currently focused on early identification of neurodevelopmental delays. 3,4 However, despite not meeting diagnostic criteria, small subclinical deficits in early neurodevelopment can have long term implications, impacting future academic success, academic productivity, and resulting in adverse health outcomes. 5,6 Additionally, early environmental insults have been hypothesized to result in subclinical deficits by destruction of substantia nigra cells which are important for later life brain function. 7-9 Beginning as early as the embryonic period, neurodevelopmental trajectories may be influenced by both biological and environmental factors. 10 Biological factors include genetic disorders, infections, perinatal brain injuries, and neuronal migration disorders. 11 Environmental factors include nutrition, levels of maternal responsiveness, and chemical exposures. 12 Biological and environmental factors interact with one another to impact neurodevelopment. 13 For example, environmental risk factors have the potential to exacerbate or improve neurodevelopment despite existing biological vulnerabilities. 14 The gestational and early life period are particularly sensitive and vulnerable to biological and environmental influences, with early insults having the ability to impact long-term neurodevelopmental trajectories. 15,16 For instance, in- utero neuronal migration disorders have been associated with motor problems while early exposures to neurotoxic chemicals may disrupt fetal brain myelination which may contribute to poor neurodevelopmental outcomes. 17 Biological and/or environmental factors may impact a single or multiple areas or domains (e.g., motor, cognitive) of neurodevelopmental functioning. 6 However, even single 10 impacts to neurodevelopmental functioning may have far reaching implications given the influence of each domain on the development of others. 18, 19 According to a book published by the US National Research Council in 2000, about 3% of developmental disabilities directly result from environmental exposures and 25% arise through gene-environment interactions. 20 However, these estimates only account for developmental disabilities and disorders, not subclinical deficits, so the true impact of chemical exposures may be higher. Environmental exposures to toxic chemicals are of particular interest since they are preventable at the policy level through regulations and restrictions. 21 Evidence suggests that anthropogenic chemicals are important neurotoxicants and may cause untreatable and permanent brain damage. 6,13,22 Prenatal and early childhood chemical exposures have been associated with intellectual disability, autism spectrum disorder, attention deficit/hyperactivity disorder (ADHD), motor delays, and learning disabilities, along with more subtle deficits such as lower IQ and subclinical attention or learning problems. 23-29 In a recent review by Bellinger et al., the total IQ loss attributed to lead, pesticides, and other neurotoxicants was comparable to, or even greater than, the loss attributed to preterm birth, traumatic brain injury, congenital heart disease, and brain tumors. 30 Furthermore, early exposures to neurotoxicants have been associated with violence, antisocial behavior, and substance abuse, resulting in additional individual and societal costs. 31-33 This is of concern given the disproportionate burden of environmental exposures to neurotoxic chemicals experienced by low-income communities and communities of color. 34-36 The neurodevelopmental effects of toxic exposures during the in-utero period of development are of primary interest given the susceptibility and vulnerability of early windows of development and the fact that the blood brain barrier has not fully matured until about 6 months after birth. 37 Additionally, it is important to note that the detrimental impacts of early environmental exposures on neurodevelopment may be further compounded by psychosocial adversity, including but not limited to stressful living conditions, substandard housing, poor nutrition, and inadequate health care. 36,38 Flame Retardant Chemicals and Neurodevelopment Chemical flame retardants are generally applied to a variety of consumer products to meet flammability regulations. 39 Historically, the flame-retardant sector was primarily dominated by chemicals 11 such polybrominated biphenyls (PBBs) and polychlorinated biphenyls (PCBs); however, evidence supporting their toxicity, environmental persistence, and adverse impacts on health subsequently led to their ban across most industrialized nations. 40,41 As a result, replacement chemicals, including polybrominated diphenyl ethers (PBDEs), rapidly became the most commonly used commercial flame retardants until the mid-2000s when they were phased out of production amid concerns regarding their bioaccumulation and toxicity to humans. 42 In recent decades, organophosphate ester (OPE) flame retardants have rapidly emerged as the most commonly produced flame retardant chemicals in the US, with a tripling in global production from 1992 to 2007. 42-44 Although previously considered less persistent in the environment than PBDEs because of their shorter half-lives, recent evidence suggests OPEs are persistent, can undergo long- range transport, and can bioaccumulate in the environment. 42,45,46 Additionally, growing toxicological evidence suggests potential toxicity to human health from environmentally relevant exposures of OPEs. 47 Previous research on historical and legacy flame retardants have found associations between prenatal exposures to these flame-retardant chemicals and adverse neurological development. 48 For instance, prenatal exposures to PCBs have been associated with adverse cognitive outcomes, such as higher risk of subclinical cognitive impairment, reduced verbal memory and language skills, and delayed mental development index, along with increased behavioral symptoms, such as ADHD, during childhood. 49-52 Similarly, prenatal PBDE exposures have been associated with adverse cognitive impacts during early childhood (including full-scale IQ and lower mental development index) and increased behavioral problems, such as externalizing problems, attention problems, poorer social skills, and more depressive symptoms. 53-57 Despite limited research similarly suggesting potential neurotoxic impacts of prenatal OPEs among children, the neurodevelopmental impacts of prenatal exposures to currently used OPEs are not as well understood. 47,48,58 Organophosphate Esters (OPEs): Properties, Environmental Prescence, and Human Exposures OPEs are a class of semi-volatile, organic chemical additives commercially applied to a wide variety of combustible consumer products and building materials to prevent or delay fires. 43,47,59 However, OPEs 12 are also commonly used as plasticizers and lubricants, contributing to their environmental ubiquity. 43,47 For the purpose of this dissertation, we focus on nine urinary OPE metabolites described in Table 1. The physical properties of OPEs, including their varying ester linkages, and their physiological properties, such as their solubility, logK ow value, vapor pressure, and bioconcentration factors (BCF), highly influence OPEs impact on the environment and exposed organisms (Table 2-3). 43,59 OPEs with higher volatility and vapor pressure, such as tri(2-chloroethyl) phosphate (TCEP), are more easily discharged into the air and deposited on dust than larger and heavier OPEs. 39,42 Chlorinated OPEs, such as TCEP, tris(1-chloro-2-propyl) phosphate (TCIPP), and tris(1,3-dichloro-2-propyl) phosphate (TDCIPP), have better water solubility while aryl and alkyl OPEs, such as Tris (2-butoxyethyl) phosphate (TBOEP), Tri-n-butyl phosphate (TNBP), Tricresyl Phosphate (TCP), and Triphenyl Phosphate (TPHP), are more hydrophobic, have similar BCFs, and generally concentrate in sediments and soils. 47,60 Non chlorinated alkyl/aryl phosphates are commonly used as plasticizers, lubricants, antifoaming agents, hydraulic fluids, and additives to lacquers while chloroalkyl phosphates are generally applied to polyurethane and other polymers for use in furniture, construction, textile, and electronic equipment. 61,62 During manufacturing, OPEs are physically incorporated into the product matrix, rather than chemically bonded with the product matrix, which, along with their semi-volatility, facilitate escape during product use and result in leakage into surrounding environments via volatilization, leaching, wear, and infiltration. 63,64 OPEs have been found in a multitude of environmental and human matrices, such as dust, the atmosphere, sediments, soil, indoor air, and various biological samples. 65-72 Humans are primarily exposed to OPEs via indoor air and dust and primary exposure routes include dermal absorption, inhalation, and ingestion. 43,47,61 There may additionally be dietary exposure of OPEs via food (i.e., fish) and drinking water, along with OPE-contaminated foodstuffs. 43 Although OPEs may be metabolized within a couple of days, OPE exposures are ubiquitous and chronic, with the daily intake of OPEs via dust ingestion among children and adults estimated to range between 0.29-64.8 and 0.07-14.9 ng/kg bw/day. 73 Recent studies suggest widespread exposure of OPEs among the U.S. population, with especially elevated OPE metabolite concentrations among pregnant 13 women, likely due in part to pregnancy related metabolic changes. 48,60,74-77 Metabolites of TDCIPP (BDCIPP) and TPHP (DPHP) are two of the most commonly detected OPEs in biomonitoring studies, with >95% detect frequencies in some cases. 47,77 Existing studies have primarily characterized factors associated with two of the most well studied OPE metabolites, DPHP and BDCIPP; however, results on concentrations by education, income, and race/ethnicity have been mixed. 77, 78 In general, women have been found to have higher concentrations of DPHP and BDCIPP metabolites than men, likely due to the use of cosmetic products, such as perfume, nail polish, and suntan lotion, which commonly contain OPEs such as TPHP. 77,79,80 Additionally, studies evaluating characteristics in pregnant mothers have found higher isopropyl-phenyl diphenyl phosphate (ip- PPP), BCEP, BDCIPP, and DPHP metabolite levels among women with higher pre-pregnancy BMI’s. 78,81 Higher household income and education have been associated with lower BDCIPP concentrations in a pregnancy cohort of women in Rhode Island, but other studies have found null differences by income and educational levels. 80,82 Residential dust and summer season have been associated with greater BDCIPP and DPHP concentrations. 78,80,82 Nulliparous pregnant women have also been found to have higher DPHP metabolite concentrations when compared to multiparous women; however, multiparous women have been associated with having higher BDCIPP concentrations when compared to nulliparous pregnant women. 78,82 14 Table 1: Organophosphate Ester (OPE) Parent Compounds and Corresponding Metabolites Analyzed OPE Parent Compounds OPE Metabolites CAS no. Chemical Name (Abbr) Chemical Structure Molecular Formula CAS no. Chemical Name (Abbr) Chemical Structure Molecular Formula 115- 86-6 Triphenyl phosphate (TPHP) C 18H 15O 4P 838- 85-7 Diphenyl Phosphate (DPHP) C 12H 11O 4P 126- 73-8 126- 71-6 Sum of Tri-n- butyl Phosphate (TNBP) and Tris(isobutyl) phosphate (TIBP) C 12H 27O 4P C 12H 27O 4P 107- 66-4 6303- 30-6 Sum of Dibutyl phosphate (DNBP) and Di-isobutyl phosphate (DIBP) C 8H 19O 4P C 8H 19O 4P 13674- 87-8 Tris(1,3-dichloro- 2- propyl)phosphate (TDCIPP) C 9H 15Cl 6O 4P 7223 6-72- 7 Bis(1,3,- dichloro-2- propyl) phosphate (BDCIPP) C 6H 11Cl 4O 4P 115- 96-8 Tri(2- chloroethyl)phosp hate (TCEP) C 6H 12Cl 3O 4P 3040- 56-0 Bis(2- chloroethyl) phosphate (BCEP) C 4H 9Cl 2O 4 P 78-51- 3 Tris(2- butoxyethyl) phosphate (TBOEP) C 18H 39O 7P 1426 0-97- 0 Bis(butoxethy l) phosphate (BBOEP) C 12H 27O 6P 13674- 84-5 Tris(1-chloro-2- propyl)phosphate (TCIPP) C 9H 18Cl 3O 4P 7894 40- 10-4 Bis(1-chloro- 2-propyl) phosphate (BCIPP) C 6H 13Cl 2O 4P 1330- 78-5 Trismethylphenyl phosphate (TMPP) ( formerly Tricresyl phosphate (TCP)) C 21H 21O 4P 3578 7-74- 7 Bis(2- methylphenyl ) phosphate (BMPP) C 14H 15O 4P 78-42- 2 Tris(2-ethylhexyl) phosphate (TEHP) C 24H 51O 4P 298- 07-7 Bis(2- ethylhexyl) phosphate (BEHP) C 16H 35O 4P 513- 08-06 Tripropyl phosphate (TPP) C 9H 21O 4P 1804- 93-9 Dipropyl Phosphate (DPRP) C 6H 15O 4P Abbr: abbreviation Data compiled using Pub Chem 15 Table 2: OPEs Physiological Properties and Uses Parent Compound Solubility (mg/L) VP (mmHG) at 25 C Logk ow BCFs Group Uses Metabol ite TPHP 1.9 1.12×10 -5 4.70 113.3 Aryl Foam seating, bedding, plastic, rubber products, nail polish.(industrial; commercial; consumer) FDA indirect additives used in food contact substances. DPHP TNBP TIBP 280 475.6 1.13×10 -3 0.013 3.82 3.60 39.81 19.51 Alkyl Alkyl Adhesives/sealants and inks/toners. (industrial; commercial; consumer) FDA indirect additives used in food contact substances. DNBP DIBP TDCIPP 7 2.98×10 -7 3.65 21.4 Chlorinated Alkyl Polyurethane foam seating and bedding (industrial; commercial; consumer) BDCIPP TCEP 7000 0.061 1.63 0.425 Chlorinated Alkyl Industrial, unspecified BCEP TBOEP 1100 1.23×10 -6 3.00 25.56 Alkyl Plasticizers, floor finishes, waxes; flame-retarding agent; adhesives and seal (industrial, commercial, consumer) BBOEP TCIPP 1200 5.64×10 -5 2.89 3.27 Chlorinated Alkyl Foam insulation, building/construction materials, foam seating, bedding products, electronic products (industrial; commercial; consumer) BCIPP TMPP (TCP) 0.3 1.1×10 -7 6.34 2534 Aryl Plastic, rubber product, lubricants, greases, and other unspecified uses. (industrial; commercial; consumer) BMPP TEHP 0.6 8.25×10 -8 9.49 3.162 Alkyl Lubricants, additives, oxidizing/reducing agents, processing aids, agricultural pesticides, plastic, rubber products BEHP TPP 6450 0.024 2.35 0.912 Aryl Plasticizer DPRP VP=Vapour pressure LogKOW: Octanol-Water Partition Coefficient BCFs= Bioconcentration Factors Data compiled using Pub Chem and Hou et al. (2016) 16 Table 3: OPE Metabolites Physical and Physiological Properties Metabolite Molecular Weight Solubility (mg/L) 25 ˚C Vapor pressure (mmHg) 25 ˚C LogK OW LogK OC LogK OA DPHP 250.187 82.38 2.26E-7 2.88 2.082 11.243 DNBP+ DIBP 210.208 210.208 430.1 574.3 3.79E-5 1.12E-4 2.29 2.14 2.180 2.027 9.049 8.899 BDCIPP 319.935 130 1.03E-7 2.18 2.640 10.998 BCEP 222.992 6456 1.22E-5 0.83 1.649 8.988 BBOEP 298.313 410.1 3.16E-7 1.74 3.775 12.116 BCIPP 251.045 * * * * * BMPP 278.240 6.652 4.69E-8 3.97 2.519 12.245 BEHP 322.421 0.05926 1.8E-7 6.07 4.235 11.845 DPRP 182.155 4115 2.39E-4 1.31 1.649 8.315 LogKOW: Octanol-Water Partition Coefficient LogKOC: Soil Adsorption Coefficient LogKOA: Octanol Air Partition Coefficient Data compiled using Wang et al. (2019) *Note: data on BCIPP’s physiological properties was not identified Prenatal OPE Exposures and Direct Biomechanisms of Neurotoxicity Detectable levels of OPEs in the chorionic villi and uterine decidua of the placenta and amniotic fluid suggest in utero transfer of OPEs to the fetus, which could directly and permanently affect neurological structures and functioning, given the time-sensitive and foundational neurological growth that occurs during the fetal period. 83-86 OPEs are hypothesized to impact neurodevelopment via disruptions to multiple neurological pathways, along with disruptions to the endocrine system, which play a critical role in early neurodevelopment. 87,88 Emerging toxicological evidence suggests that OPEs may result in perturbations to various neurotransmitters which play a foundational role in healthy brain functioning. 89-91 Neurotransmitters are exogenous chemicals important for neuronal communication throughout the body and involved in a variety of critical neurological processes, including neurotransmission, differentiation, neuronal growth, and the development of neural circuitry. 92 Prior animal studies in mice have found disruptions to gamma- aminobutyric acid (GABA), glutamate, dopamine, and serotonin from OPE exposures, along with increased ambulatory behavior and impaired learning and memory. 89,93-96 Additionally, animal studies on zebrafish suggest that OPEs may impact histamine levels and adversely impact motor development and alter behavior 17 in manners which persist into adulthood. 97-100 Since neurotransmitters play such an important neurological role, any OPE related disruptions may adversely impact essential neurological mechanisms for healthy neurodevelopment. 92 For instance, dopamine plays a critical role in motor control, learning, emotional, and executive functioning, and serotonin modulates neural activity for a variety of neuropsychological processes, with disturbances in these neurotransmitters implicated in a variety of psychiatric and neurodegenerative diseases. 101,102 GABA is an important inhibitory neurotransmitter, while glutamate is an excitatory neurotransmitter which serves as a primary mediator of nervous system plasticity. 92 There is also evidence that OPEs may adversely impact early neurodevelopment via disruptions to various neuronal processes, including cell differentiation and proliferation and synaptogenesis and network formation. 90,91,97,103 In vitro studies using undifferentiated and differentiating PC12 cells have found concentration-dependent neurotoxicity between OPEs and decreased cell number and altered neurodifferentiation, with neurotoxic impacts of some OPE analytes equivalent to or greater in equimolar concentration than chlorpyrifos (CPF). 104 In particular, TDCIPP and tris (2,3-dibromopropyl) phosphate (TDBPP) have been found in toxicological studies using PC12 cells to promote differentiation in both neuronal phenotypes while TCEP and TCPP have been found to promote only the cholinergic phenotype. 104 Additionally, adverse effects were observed in both the undifferentiated and neurodifferentiation stage, strongly suggesting that the developing nervous system is particularly vulnerable to OPEs during the earliest stages of neural formation. 104 Other in vitro assays suggest that OPEs and PBDEs have similar potential for neurodevelopmental toxicity via impacts on neuronal proliferation, neurite outgrowth, synaptogenesis, and network formation, all processes critical to neurodevelopment. 97,105 OPEs are also hypothesized to potentially contribute to neurotoxicity via neuroinflammatory disruptions and direct lesions to critical brain structures. 106-108 In vitro and in vivo studies in mice have found increased cytokine gene and receptor expression suggesting OPEs exposures may induce inflammatory response and result in neuronal damage through microglia-mediated inflammation as well as disturbances to synaptogenesis. 91,106,109 Animal studies in zebrafish and freshwater clams have additionally found that OPEs may increase reactive oxygen species. 110,111 Previous observational research further 18 suggests there may be an association between elevations in biomarkers of oxidative stress and neurodevelopmental disorders across the lifespan. 112-116 Animal studies in rats also suggest that acute and intermediate duration exposure of TCEP may result in brain lesions in the hippocampus and in the cerebral cortex and chronic-duration exposure may result in brain stem lesions, which may contribute to neurodevelopmental conditions. 107,108,117 Along with neurologically mediated pathways, OPEs may impact neurological health through mechanisms involving the endocrine system, with growing evidence of potential sex-specific toxicity. 118,119 The endocrine system plays a crucial role in healthy fetal brain development, with any disruptions to the endocrine system having the potential to result in neurodevelopmental damage. 88 Both experimental and observational evidence suggests disruptions to thyroid and sex hormones from exposures to OPEs, with potential alterations to thyroid-stimulating hormone (TSH), triiodothyronine (T3), thyroxine (T4) levels and the hypothalamic-pituitary-gonadal (HPG), hypothalamic-pituitary-interrenal (HPI), and hypothalamic-pituitary-thyroid (HPT) axis, along with disruptions to sex steroids and sex steroid binding globulins. 119-125 For instance, a study published by Tao et al. in 102 mother-newborn pairs in Wuhan, China found associations between multiple urinary OPE metabolites measured across pregnancy and disruptions to TSH levels in newborns (measured via time-resolved immunofluorescence assay), with evidence of sex- specific specific effects. 126 Altogether, existing experimental evidence suggests that there are various biologically plausible mechanisms through which prenatal OPE exposures may directly impact early neurodevelopment. Existing evidence further supports that such effects could be sex specific. Prenatal OPE Exposures and Indirect Biomechanisms of Neurotoxicity Prenatal OPE exposures may also impact neurodevelopment through indirect, placenta-mediated pathways, such as through disruptions to the maternal endocrine system. 127 Fetal formation of vital endocrine structures begins during the early in-utero period, with high reliance on placentally transferred maternal thyroid hormones during early pregnancy, and less reliance across mid-pregnancy as fetal production of thyroid hormone begins. 128 Exogenous exposures to endocrine disrupting chemicals can 19 impact the homeostasis of maternal thyroid levels, with subclinical thyroid disruptions potentially resulting in functional and structural brain changes that impair healthy child neurodevelopment. 129,130 There is further evidence of sex-specific, placenta-mediated associations between prenatal OPE exposures, thyroid levels, and neurotoxicity, with animal studies in rats suggesting sex-specific accumulation of OPEs in the placenta and impacts on neurotransmitter production and serotonergic projections in fetal forebrains, specifically among exposed males. 96,131 This is consistent with prior research which suggests that the sex-specific accumulation of toxic chemicals in the placenta may differentially influence thyroid functioning, due to the hypothesized sex differences in chemical uptake via thyroid hormone transporting membrane proteins in placental tissue, impacting maternal supply of thyroid hormones to support fetal development. 132 Prior studies on brominated flame retardants have observed similar potential mechanisms, with higher concentrations of PBDEs and thyroid levels among placenta samples of male infants, despite similar maternal serum concentrations across sex. 132 Additionally, a previous observational study found increased risk of ADHD among children with higher prenatal OPE exposures partially mediated by maternal total triiodothyronine to total thyroxine ratio. 133 Since fetal and maternal thyroid hormones continuously interact with one another during pregnancy, playing a critical role in the developing central nervous system, it is likely that both direct and indirect insults to the fetal endocrine system from prenatal OPE exposures simultaneously contribute to neurtoxocity. 134 Birth Outcomes as Early Indicators of Later Neurodevelopmental Outcomes Early birth outcomes, including gestational age (GA) at birth and birthweight, are strongly associated with later neurodevelopmental outcomes. 135 For instance, school-age children born very preterm are 35% to 50% more likely to experience neurodevelopmental impairments in the US, including but not limited to, deficits to cognitive, language, fine and gross motor, processing speed, and executing functioning skills. 135,136 Prior research further suggests that earlier GA at birth is a risk factor for neurodevelopmental delays, with further evidence of improved cognitive and psychomotor performance with longer gestational age at birth. 137-139 Independent of preterm birth, lower birthweight has also been found to be a powerful 20 predictor of adverse neurodevelopmental outcomes. 140 However, GA at birth and birthweight are intrinsically linked, with lower birthweight often found among infants born at an earlier GA at birth and fetal growth restriction common among preterm infants, making it difficult to separate risk factors of each birth outcome from one another. 140 The strong associations observed between birth outcomes and later neurodevelopmental outcomes are hypothesized to result from a multitude of reasons, including the many shared mechanisms critical to both in-utero growth and neurological development. 140,141 In other words, the adverse effects of toxic prenatal insults which increase neurodevelopmental susceptibility to disorders may also impair fetal growth and heighten risk of early GA at birth. 95 As a result, early birth outcomes can potentially serve as early indicators of increased risk to later adverse neurodevelopmental outcomes. Prenatal OPE Exposures and Biomechanisms Contributing to Adverse Birth Outcomes Growing toxicological evidence suggests OPEs adversely impact birth outcomes via similar mechanisms as they impact neurotoxicity, including endocrine disrupting pathways, oxidative stress, and placental disruptions. 118,120,142 Since thyroid hormones, the hypothalamic pituitary-adrenal (HPA), and hypothalamic pituitary gonadal (HPG) axis also play a vital role in fetal growth, any potential OPE related disruptions could adversely impact infant birth outcomes. 125,143-145 OPEs have been found to result in oxidative stress in zebrafish, resulting in concern since increased reactive oxygen species have been associated with negative reproductive outcomes. 110,111,146,147 Epidemiological studies further suggest OPEs exposures may result in placenta mediated pregnancy complications. 83 Placental disruptions and pregnancy complications have been associated with adverse infant birth outcomes. 148,149 Sex-specific impacts on birth outcomes from OPE exposures are also hypothesized, given the estrogen-like effects of OPEs acting on various hormone related pathways and sex differences in estrogen receptor expression. 143,150,151 Since OPEs have different nuclear receptor properties, such as estrogen receptor agonistic activity, androgen receptor antagonistic activity, glucocorticoid receptor antagonistic activity, and pregnane X receptor agonistic activity, they could simulate or antagonize the binding of related hormones to receptors, directly impacting the endocrine system in sex-specific ways. 125,143,152 Additional 21 studies suggest possible growth inhibition via down regulation of growth hormone/insulin-like growth factor (GH/IGF) genes. 153,154 Existing experimental evidence supports the potential for OPEs exposures to adversely impact offspring birthweight and body length in animal studies. 155 A study by Farhat et al. published in 2013 found that TDCIPP exposed chicken embryos had significantly decreased weight at hatching and a suggestive delay in pipping time, however this association did not reach statistical significance. 122 A second study by Chen et al. published in 2015 found that oral exposures to TPP and TCEP of about 300 mg/kg for 35 days in adult mice resulted in decreased body, liver, and testis weights. 156 Yu et al. published a study in 2017 which found that TDCIPP exposure of 580 or 7500 ng TDCIPP/L for 240 days could significantly inhibit offspring body length and mass of zebrafish. 157 Additional studies on zebrafish have found associations between TBOEP and reduced egg production and length in zebrafish and TBOEP and TDCIPP with reduced body weight and body length in female zebrafish. 153,154,158,159 Epidemiological Evidence of Early Developmental Impacts Neurodevelopment Epidemiological evidence evaluating the association between prenatal OPEs and early neurodevelopment in human populations is limited; however evidence is generally suggestive of adverse associations between OPE exposures and neurodevelopmental outcomes. 160-162 A study by Castorina et al. among the CHAMACOS cohort comprised of all Hispanic participants with at most a high school education (~79%), evaluated the association between prenatal OPE metabolites measured during a second prenatal study visit (mean gestation age at collection=26.0 (2.4) weeks) and children’s neurodevelopment at 7 years old, using the Wechsler Intelligence Scale for Children (WISC-IV), Behavior Assessment System for Children 2 (BASC-2), and Conners’ ADHD/DSM-IV Scales (CADs). 160 For each ten-fold increase in prenatal urinary concentrations of DPHP, this study found decreases of 2.9 points (95% CI: -6.3, 0.5) and 3.9 points (95% CI: -7.3, -0.5) in the Full-Scale intelligence quotient and Working Memory, respectively. 160 Additionally, urinary prenatal isopropyl-phenyl phenyl phosphate (ip-PPP) concentrations were associated 22 with increased hyperactivity in 7 year old children when using maternal informed BASC-2 measures. Another study by Doherty et al. among the Pregnancy, Infection, and Nutrition Study (PIN) participants, evaluated OPEs in spot prenatal maternal urine measured at approximately 27 weeks’ gestation and child cognitive development at two and three years old. 161 This study found that higher ip-PPP concentrations were associated with lower scores on the Mullen Scales of Early Learning (MSEL) Cognitive composite score (𝛽 : -2.61, 95% CI: -5.69, 0.46), the Fine Motor Scale (𝛽 : -3.08, 95% CI: -5.26, -0.91), and the Expressive Language Scale (𝛽 : -1.21, 95% CI: -2.91, 0.49). Additionally, prenatal ip-PPP concentrations were inversely associated with age-standardized scores on the MacArthur-Bates Communicative Development Inventories (MB-CDI) Vocabulary assessment (𝛽 :-1.19, 95% CI: -2.53, 0.16). 161 Doherty et al. published another study on the PIN cohort which found a positive association between higher prenatal BDCIPP concentrations and higher scores on the Behavioral Symptoms Index (1 st vs. 3 rd tertile 𝛽 : 3.03, 95% CI: 0.40, 5.67) and a non-significant positive association with externalizing problems scores (1 st vs. 3 rd tertile 𝛽 : 2.49, 95% CI: -0.12, 5.10) at 36 months. However, this study also found an inverse significant association between higher concentrations of ip-PPP and internalizing problems composite scores (1 st vs. 3 rd tertile 𝛽 : -3.74, 95% CI: -6.75, -0.74). 162 Another study by Liu et al. among 184 pregnant women in Wuhan, China found that prenatal BDCIPP concentrations and average molar concentrations of chlorinated- alkyl OPEs were associated with decreased Psychomotor Development Index (PDI) (BDCIPP 𝛽 = -3.50, 95% CI: -5.86, -1.14 and chlorinated-alkyl 𝛽 = -3.24, -5.95, -0.53) and decreased Mental Development Index (BDCIPP 𝛽 = -5.75, 95% CI: -8.94, -2.55 and chlorinated-alkyl 𝛽 = -5.86, 95% CI: -9.52, -2.20) in 2 year old children, specifically among boys. 163 Another study by Percy et al. on the HOME cohort found possible patterns of effect modification by socio-economic status for the association between postnatal urinary OPEs measured at ages 1-5 years old and children’s cognition at 8 years old, with each log-unit increase in BDCIPP, BCEP, and DPHP associated with 1 to 2 point decreases in the Full Scale IQ among children with lower maternal education, non-white race, lower income, or in more deprived neighborhoods, and similar patterns across the Perceptual Reasoning, Verbal Comprehension, and Working Memory Index 23 scores. 164 However, there were no significant associations between urinary OPE metabolites and cognitive abilities at 8 or 12 years old. Two other studies on the Norwegian Mother, Father, and Child Cohort Study (MoBa) have investigated associations between prenatal OPEs and 18-month-old language skills and clinically-assessed ADHD at 3 years old, finding that higher prenatal DNBP concentrations were associated with 1.71 times (95% CI: 1.13, 2.58) the odds of having ADHD at 3 years old but they found no statistically significant association between prenatal OPEs and language skills at 18 months. 165,166 Limited studies have additionally evaluated OPEs in dust or passive sampler concentrations during early childhood and children’s neurodevelopment. 167,168 A study by Sugeng et al. among 42 children participating in the Linking maternal Nutrition to Child health (LINC) evaluated the association between OPE concentrations in house dust and hand wipes (measured at median age of 13 months) and children’s behavior at a median age of 18 months. 167 This study found that a 10-fold increase in TCEP dust concentrations were associated with a 13 point increase in the externalizing problems scale (p=0.043). 167 Additionally, 10-fold increases in bisphenol A bis(diphenyl phosphate) (BPA-BDPP) and resorcinol bis(diphenyl phosphate) (PBDPP) in household dust were associated with 4.1 and 2.6 point increases on the internalizing problem scales (p=0.004, p=0.042 respectively) and increased BPA-BDPP was also associated with elevated scores on the externalizing problems scale (7.6 points, p=0.015) and total behavioral problems scale (16 points, p=0.007) at 18 months. 167 A cross-sectional study published by Lipscomb et al. found that 3-5 year old children who attended preschool (N=72) and had higher OPE levels in silicone passive samplers worn for 7 days had less responsible behavior (p=0.07) and more externalizing behavior problems (p=0.03). 168 Another cross-sectional study by Hutter et al. found associations between higher concentrations of TCEP in PM10 and PM2.5 dust samples in schools and decreased cognitive performance among children 6-8 years old participating in Austria. 169 Epidemiologic studies evaluating the association between OPE mixtures and neurodevelopment have also been limited. 133 A recent study by Choi et al. on 295 children with ADHD and 555 children without ADHD from the Norwegian Patient Registry found that a quartile decrease of OPE-phthalate mixtures, including DPHP, DNBP, and 6 phthalate metabolites (collected at 17 weeks’ gestation), resulted 24 in an ADHD risk ratio of 0.68 (0.64, 0.72). 133 Another study by Percy et al. additionally evaluated associations between OPE concentrations both independently and in mixtures across prenatal (16 and 26 weeks of GA and delivery) and postnatal (1, 2, 3, and 5 years) urinary samples and neurobehavioral symptoms at 3 and 8 years using the BASC-2. 170 Prenatal OPE metabolites using latent variable values at 16 weeks were associated with lower externalizing (𝛽 =-5.74; 95% CI: -11.24, -0.24) and total behavioral problems (𝛽 = -5.25; 95% CI: -10.33, -0.19) among 3-year-old children, while higher concentrations at delivery were associated with higher overall behavioral problems (𝛽 = 2.87; 95% CI: 0.13, 5.61) at 3 years old. Higher OPE values at 3 years old were also associated with lower externalizing behaviors at 8 years old (𝛽 = -2.62; 95% CI: -5.13, -0.12). Overall, OPE mixtures using quantile g-computations were consistent with latent variable analysis results; however, results were mixed across timepoints. While the majority of studies have shown higher OPE concentrations to be associated with reduced cognitive development, another study by Percy et al. among the Health Outcomes and Measures of the Environment (HOME) study, found no associations between urinary OPE metabolite mixtures and child cognitive measures but found that prenatal urinary BCEP concentrations were associated with a small increase in 8 year old children’s full-scale IQ at 8 years old (𝛽 = 0.81, 95% CI: 0.00, 1.61). 171 Birth Outcomes Limited epidemiological research has examined the association between prenatal OPE exposures and infant birth outcomes, but overall, evidence suggests sex-specific adverse impacts from OPE exposures on GA at birth and birthweight. 172 Hoffman et al. published a study among a cohort of 349 predominately white (~79.7%) and educated (~69.6%) women residing in North Carolina and participating in the Pregnancy Infection and Nutrition Study (PIN) between 2002 through 2005. 173 This study examined the association between prenatal BDCIPP, DPHP, isopropyl-phenyl phenyl phosphate (ip-PPP), and bis(1- chloro-2-propyl) 1-hydroxy-2-proyl phosphate (BCIPHIPP) metabolites (gestational age at collection: 24- 30 weeks) on infant birthweight and gestational length, finding significantly increased odds of preterm delivery with higher prenatal BDCIPP levels (OR: 3.99, 95% CI: 1.08, 14.78) and earlier delivery of about 25 1 week for higher ip-PPP concentrations (95% CI: -1.9, -0.2) among female infants only. 173 However, higher prenatal ip-PPP levels were associated with decreased odds of preterm birth among male infants (OR: 0.21, 95% CI: 0.06, 0.68), with overall results indicating sex-specific impacts of OPEs on infant birth outcomes. 173 Similarly, a case-control study published by Luo et al. (cases N=113, controls N=226) among women residing in Wuhan, China from 2014 to 2016 found that higher prenatal DPHP levels were associated with significantly increased risk of low birthweight (OR: 4.62, 95% CI: 1.72, 12.40), with a significant dose-response relationship (p-trend <0.01). 174 After stratification by infant sex, the association was only observed among females, further suggesting sex-specific OPE impacts. 174 Another study by Luo et al. among a pregnancy cohort of 213 primarily college educated (~81.2%) women residing in Wuhan, China and recruited from 2014-2016 evaluated OPEs concentrations across pregnancy trimesters and infant birthweight. 175 This study found that higher BDCIPP (𝛽 : -85.12, 95% CI: -148.98, -21.27) and bis(2- butoxyethyl) phosphate (BBOEP) (𝛽 : -35.48, 95% CI: -68.23, -2.73) in the third trimester, 4- hydroxyphenyl-diphenyl phosphate (4-HO-DPHP) (𝛽 : -48.46, 95% CI: -94.50, -2.42) in the second trimester, and DPHP (𝛽 : -43.34, 95% CI: -81.35, -5.33) in the first trimester were negatively associated with birthweight. 175 After sex stratification, third trimester BBOEP (𝛽 : -48.11, 95% CI: -90.20, -6.01), and BDCIPP (𝛽 : -89.62, 95% CI: -165.04, -14.21) were only significant among males and first trimester DPHP was only significant among females (𝛽 : -59.09, 95% CI: -114.96, -3.23). 175 Crawford et al. published a study on 56 primarily college educated (45.0%) and non-Hispanic white (64.0%) women affiliated with the Women & Infants Hospital Rhode Island (WIHRI) in 2014 with urine samples collected throughout three pregnancy timepoints. 176 This study found BDCIPP was associated with increased infant length (𝛽 : 0.44 cm, 95% CI: 0.01, 0.87) and weight in males (𝛽 : 0.14 kg, 95% CI: 0.03, 0.24) but DPHP was inversely associated with overall infant abdominal circumference (𝛽 : -0.50 cm, 95% CI: -0.86, -0.14) and female weight (𝛽 :-0.19 kg, 95% CI: -0.36, -0.02). 176 There have additionally been a limited number of studies evaluating the impacts of OPE mixtures on infant birth outcomes. 177,178 A cohort study (n=76) by Kuiper et al. on primarily white (~53%) and college 26 educated women (~60%) enrolled from 2017-2019 in the ORigins of Child Health And Resilience in Development (ORCHARD) pregnancy cohort in Baltimore, MD found no significant associations between single OPE metabolites or OPE metabolite mixtures and infant birthweight for gestational age z-scores or gestational age at delivery. 177 However, sex-specific models were not evaluated. 177 Bommarito et al. published another case-control study among a subset of LIFECODE study pregnancy cohort participants recruited in Boston, MA in 2006 with urine samples collected three times throughout pregnancy (cases= 31 small-for-gestational age (SGA) and 28 large-for-gestational age (LGA), controls= 31 appropriate for gestational age). This study used adjusted multinomial logistic regression models and quantile g- computation to explore both single and mixtures effects of OPE metabolites on birthweight outcomes. In single metabolite associations, this study found lower odds of LGA with an IQR increase in DPHP (OR: 0.40, 95% CI: 0.18, 0.87). 178 In quantile g-computation models, lower odds of LGA birth were estimated for higher OPE concentrations mixtures of DPHP and BDCIPP (OR: 0.49, 95% CI: 0.27, 0.89). 178 Additionally, closer inspection of quantile g-computation weights indicated stronger contributions of DPHP in the observed association when compared to BDCIPP. Another study by Yang et al. among the Health Outcomes and Measures of Environment (HOME) study examined associations between prenatal OPE metabolite concentrations at 16 weeks, 26 weeks, and delivery and children’s anthropometric measures, ponderal index, and weekly growth rates at 4 weeks old, finding positive associations between DPHP at 16 weeks and BCEP and BDCIPP at delivery and weight, length, and head circumference z-scores. After sex- stratification, positive associations between DPHP at 16-weeks and BDCIPP at delivery and anthropometric measures were observed among males only and positive associations between BCEP at delivery and anthropometric measures observed among females only, with consistent results in OPE metabolite mixtures models examined using Bayesian inference procedure for lagged kernel machine regression (MFVB- LKMR). 179 Current Gaps in the Literature on OPEs and Neurodevelopment As summarized in this chapter, existing toxicological and epidemiological evidence suggests ubiquitous OPE exposures which may pose a neurotoxic threat to early neurological development (Figure 27 1). However, there currently exists various gaps in the literature examining the impacts of OPEs on early neurodevelopment. For one, most of the existing OPE literature has focused on in-vitro and animal studies. 47 Thus, there is limited epidemiological research which has assessed the impacts of OPEs on infant birth outcomes and neurodevelopment in human populations. 74,168,173-175 Additionally, among the limited studies which have explored these associations, few have investigated these questions among lower income and diverse populations. Studies have primarily focused on evaluating the impacts of two OPE metabolites, BDCIPP and DPHP, resulting in a lack of evidence on the impacts of other OPE metabolites on infant birth outcomes and neurodevelopment. Moreover, most of the literature has focused on single metabolite models rather than examining OPE mixtures on infant birth outcomes and neurodevelopment. Since OPE exposures are likely to co-occur and chemicals may differ in their neurodevelopmental impacts, with possible additive or synergistic effects, exposures to mixtures of OPEs, in addition to effects of single chemicals, are critical to understanding the prenatal impacts of these everyday exposures on early neurodevelopment. This dissertation will address many of the previously mentioned gaps in literature. Along with adding to the limited number of epidemiological studies in human populations on OPEs and infant birth outcomes and neurodevelopment, this dissertation will be evaluating the impacts of both single metabolite and OPE mixtures on neurodevelopmental outcomes of interest. Additionally, nine OPE metabolites, many of which have not been studied, will be evaluated in each of the three Study Projects. Lastly, this study will explore these associations among a pregnancy cohort of predominantly Hispanic women. 28 Figure 1: OPE Exposures, Mechanisms, and Existing Observational Evidence 29 REFERENCES 1. Zablotsky B, Black LI, Maenner MJ, et al. Prevalence and Trends of Developmental Disabilities among Children in the United States: 2009–2017. Pediatrics 2019; 144: e20190811. DOI: 10.1542/peds.2019-0811. 2. Boyle CA, Boulet S, Schieve LA, et al. Trends in the prevalence of developmental disabilities in US children, 1997-2008. Pediatrics 2011; 127: 1034-1042. 2011/05/25. DOI: 10.1542/peds.2010-2989. 3. Trasande L and Liu Y. Reducing The Staggering Costs Of Environmental Disease In Children, Estimated At $76.6 Billion In 2008. Health Affairs 2011; 30: 863-870. DOI: 10.1377/hlthaff.2010.1239. 4. Prevention CfDCa. CDC’s Work on Developmental Disabilities, (2021, accessed 01/24 2022). 5. Rauh VA and Margolis AE. Research Review: Environmental exposures, neurodevelopment, and child mental health - new paradigms for the study of brain and behavioral effects. J Child Psychol Psychiatry 2016; 57: 775-793. 2016/03/14. DOI: 10.1111/jcpp.12537. 6. Grandjean P and Landrigan PJ. Neurobehavioural effects of developmental toxicity. The Lancet Neurology 2014; 13: 330-338. DOI: 10.1016/S1474-4422(13)70278-3. 7. Grandjean P. Effects on reserve capacity: significance for exposure limits. Science of the total environment 1991; 101: 25-32. 8. Landrigan PJ, Sonawane B, Butler RN, et al. Early environmental origins of neurodegenerative disease in later life. Environ Health Perspect 2005; 113: 1230-1233. 9. Grandjean P and Landrigan PJ. Developmental neurotoxicity of industrial chemicals. Lancet 2006; 368: 2167-2178. 2006/12/19. DOI: 10.1016/s0140-6736(06)69665-7. 10. Council NR. From neurons to neighborhoods: The science of early childhood development. 2000. 11. Villagomez AN, Muñoz FM, Peterson RL, et al. Neurodevelopmental delay: Case definition & guidelines for data collection, analysis, and presentation of immunization safety data. Vaccine 2019; 37: 7623-7641. DOI: 10.1016/j.vaccine.2019.05.027. 12. Stillerman KP, Mattison DR, Giudice LC, et al. Environmental exposures and adverse pregnancy outcomes: a review of the science. Reprod Sci 2008; 15: 631-650. 2008/10/07. DOI: 10.1177/1933719108322436. 13. Miguel PM, Pereira LO, Silveira PP, et al. Early environmental influences on the development of children's brain structure and function. Dev Med Child Neurol 2019; 61: 1127-1133. 2019/02/12. DOI: 10.1111/dmcn.14182. 14. Walker SP, Wachs TD, Grantham-McGregor S, et al. Inequality in early childhood: risk and protective factors for early child development. The Lancet 2011; 378: 1325-1338. DOI: https://doi.org/10.1016/S0140-6736(11)60555-2. 15. Bateson P, Barker D, Clutton-Brock T, et al. Developmental plasticity and human health. Nature 2004; 430: 419-421. DOI: 10.1038/nature02725. 16. Payne-Sturges DC, Cory-Slechta DA, Puett RC, et al. Defining and Intervening on Cumulative Environmental Neurodevelopmental Risks: Introducing a Complex Systems Approach. Environ Health Perspect 2021; 129: 35001-35001. 2021/03/10. DOI: 10.1289/EHP7333. 17. Monk C and Hane A. Fetal and Infant Brain–Behavior Development: Milestones & Environmental Influences. 2014. 18. Lingam R, Golding J, Jongmans MJ, et al. The association between developmental coordination disorder and other developmental traits. Pediatrics 2010; 126: e1109-1118. 2010/10/20. DOI: 10.1542/peds.2009-2789. 19. Lingam R, Jongmans MJ, Ellis M, et al. Mental health difficulties in children with developmental coordination disorder. Pediatrics 2012; 129: e882-891. 2012/03/28. DOI: 10.1542/peds.2011-1556. 20. Council NR. Scientific frontiers in developmental toxicology and risk assessment. 2000. 21. Rojas-Rueda D, Morales-Zamora E, Alsufyani WA, et al. Environmental Risk Factors and Health: An Umbrella Review of Meta-Analyses. Int J Environ Res Public Health 2021; 18: 704. DOI: 10.3390/ijerph18020704. 30 22. Slotkin TA and Seidler FJ. Does mechanism matter? Unrelated neurotoxicants converge on cell cycle and apoptosis during neurodifferentiation. Neurotoxicology and teratology 2012; 34: 395-402. 23. Landrigan PJ, Lambertini L and Birnbaum LS. A research strategy to discover the environmental causes of autism and neurodevelopmental disabilities. Environ Health Perspect 2012; 120: a258-a260. 24. Boucher O, Jacobson SW, Plusquellec P, et al. Prenatal methylmercury, postnatal lead exposure, and evidence of attention deficit/hyperactivity disorder among Inuit children in Arctic Quebec. Environ Health Perspect 2012; 120: 1456-1461. 25. Froehlich TE, Anixt JS, Loe IM, et al. Update on environmental risk factors for attention- deficit/hyperactivity disorder. Current psychiatry reports 2011; 13: 333-344. 26. Roze E, Meijer L, Bakker A, et al. Prenatal exposure to organohalogens, including brominated flame retardants, influences motor, cognitive, and behavioral performance at school age. Environ Health Perspect 2009; 117: 1953-1958. 2010/01/06. DOI: 10.1289/ehp.0901015. 27. Kofman O, Berger A, Massarwa A, et al. Motor inhibition and learning impairments in school- aged children following exposure to organophosphate pesticides in infancy. Pediatric research 2006; 60: 88-92. 28. Rauh V, Arunajadai S, Horton M, et al. Seven-Year Neurodevelopmental Scores and Prenatal Exposure to Chlorpyrifos, a Common Agricultural Pesticide. Environ Health Perspect 2011; 119: 1196- 1201. DOI: doi:10.1289/ehp.1003160. 29. Cho K, Frijters JC, Zhang H, et al. Prenatal exposure to nicotine and impaired reading performance. J Pediatr 2013; 162: 713-718. e712. 30. Bellinger DC. Interpreting epidemiologic studies of developmental neurotoxicity: conceptual and analytic issues. Neurotoxicol Teratol 2009; 31: 267-274. 2009/07/15. DOI: 10.1016/j.ntt.2009.06.006. 31. Schwartz J. Societal benefits of reducing lead exposure. Environmental Research 1994; 66: 105- 124. 32. Nevin R. Understanding international crime trends: the legacy of preschool lead exposure. Environmental research 2007; 104: 315-336. 33. Gould E. Childhood lead poisoning: conservative estimates of the social and economic benefits of lead hazard control. Environ Health Perspect 2009; 117: 1162-1167. 34. Morello-Frosch R and Shenassa ED. The environmental "riskscape" and social inequality: implications for explaining maternal and child health disparities. Environ Health Perspect 2006; 114: 1150-1153. DOI: 10.1289/ehp.8930. 35. Cory‐Slechta DA, Virgolini MB, Rossi‐George A, et al. Lifetime consequences of combined maternal lead and stress. Basic & clinical pharmacology & toxicology 2008; 102: 218-227. 36. Rauh VA, Landrigan PJ and Claudio L. Housing and health: intersection of poverty and environmental exposures. Annals of the New York Academy of Sciences 2008; 1136: 276-288. 37. Adinolfi M. The development of the human blood‐CSF‐brain barrier. Developmental Medicine & Child Neurology 1985; 27: 532-537. 38. Mohai P, Lantz PM, Morenoff J, et al. Racial and socioeconomic disparities in residential proximity to polluting industrial facilities: evidence from the Americans' Changing Lives Study. American journal of public health 2009; 99: S649-S656. 39. Stapleton HM, Klosterhaus S, Eagle S, et al. Detection of organophosphate flame retardants in furniture foam and U.S. house dust. Environ Sci Technol 2009; 43: 7490-7495. 2009/10/24. DOI: 10.1021/es9014019. 40. Pijnenburg AMCM, Everts JW, de Boer J, et al. Polybrominated Biphenyl and Diphenylether Flame Retardants: Analysis, Toxicity, and Environmental Occurrence. In: Ware GW and Gunther FA (eds) Reviews of Environmental Contamination and Toxicology: Continuation of Residue Reviews. New York, NY: Springer New York, 1995, pp.1-26. 41. Kimbrough RD. Polychlorinated Biphenyls (PCBs) and Human Health: An Update. Critical Reviews in Toxicology 1995; 25: 133-163. DOI: 10.3109/10408449509021611. 31 42. Blum A, Behl M, Birnbaum L, et al. Organophosphate Ester Flame Retardants: Are They a Regrettable Substitution for Polybrominated Diphenyl Ethers? Environ Sci Technol Lett 2019; 6: 638- 649. 2020/06/05. DOI: 10.1021/acs.estlett.9b00582. 43. Hou R, Xu Y and Wang Z. Review of OPFRs in animals and humans: Absorption, bioaccumulation, metabolism, and internal exposure research. Chemosphere 2016; 153: 78-90. DOI: https://doi.org/10.1016/j.chemosphere.2016.03.003. 44. Greaves AK and Letcher RJ. A Review of Organophosphate Esters in the Environment from Biological Effects to Distribution and Fate. Bull Environ Contam Toxicol 2017; 98: 2-7. 2016/08/12. DOI: 10.1007/s00128-016-1898-0. 45. Sühring R, Scheringer M, Rodgers TFM, et al. Evaluation of the OECD POV and LRTP screening tool for estimating the long-range transport of organophosphate esters. Environmental Science: Processes & Impacts 2020; 22: 207-216. 10.1039/C9EM00410F. DOI: 10.1039/C9EM00410F. 46. Zhang X, Sühring R, Serodio D, et al. Novel flame retardants: Estimating the physical-chemical properties and environmental fate of 94 halogenated and organophosphate PBDE replacements. Chemosphere 2016; 144: 2401-2407. 2015/11/28. DOI: 10.1016/j.chemosphere.2015.11.017. 47. Yang J, Zhao Y, Li M, et al. A Review of a Class of Emerging Contaminants: The Classification, Distribution, Intensity of Consumption, Synthesis Routes, Environmental Effects and Expectation of Pollution Abatement to Organophosphate Flame Retardants (OPFRs). Int J Mol Sci 2019; 20 2019/06/20. DOI: 10.3390/ijms20122874. 48. Vuong AM, Yolton K, Cecil KM, et al. Flame retardants and neurodevelopment: An updated review of epidemiological literature. Curr Epidemiol Rep 2020; 7: 220-236. 2021/01/08. DOI: 10.1007/s40471-020-00256-z. 49. Berghuis SA, Van Braeckel KNJA, Sauer PJJ, et al. Prenatal exposure to persistent organic pollutants and cognition and motor performance in adolescence. Environ Int 2018; 121: 13-22. DOI: https://doi.org/10.1016/j.envint.2018.08.030. 50. Caspersen IH, Aase H, Biele G, et al. The influence of maternal dietary exposure to dioxins and PCBs during pregnancy on ADHD symptoms and cognitive functions in Norwegian preschool children. Environ Int 2016; 94: 649-660. DOI: https://doi.org/10.1016/j.envint.2016.06.033. 51. Ruel MVM, Bos AF, Soechitram SD, et al. Prenatal exposure to organohalogen compounds and children’s mental and motor development at 18 and 30 months of age. NeuroToxicology 2019; 72: 6-14. DOI: https://doi.org/10.1016/j.neuro.2019.01.003. 52. Verner M-A, Hart JE, Sagiv SK, et al. Measured Prenatal and Estimated Postnatal Levels of Polychlorinated Biphenyls (PCBs) and ADHD-Related Behaviors in 8-Year-Old Children. Environ Health Perspect 2015; 123: 888-894. DOI: doi:10.1289/ehp.1408084. 53. Braun JM, Yolton K, Stacy SL, et al. Prenatal environmental chemical exposures and longitudinal patterns of child neurobehavior. NeuroToxicology 2017; 62: 192-199. DOI: https://doi.org/10.1016/j.neuro.2017.07.027. 54. Zhang H, Yolton K, Webster GM, et al. Prenatal PBDE and PCB Exposures and Reading, Cognition, and Externalizing Behavior in Children. Environ Health Perspect 2017; 125: 746-752. DOI: doi:10.1289/EHP478. 55. Vuong AM, Yolton K, Xie C, et al. Childhood polybrominated diphenyl ether (PBDE) exposure and neurobehavior in children at 8 years. Environmental Research 2017; 158: 677-684. DOI: https://doi.org/10.1016/j.envres.2017.07.028. 56. Kim S, Eom S, Kim H-J, et al. Association between maternal exposure to major phthalates, heavy metals, and persistent organic pollutants, and the neurodevelopmental performances of their children at 1 to 2years of age- CHECK cohort study. Science of The Total Environment 2018; 624: 377-384. DOI: https://doi.org/10.1016/j.scitotenv.2017.12.058. 57. Wang B-L, Pang S-t, Zhang X-l, et al. Levels of Polybrominated Diphenyl Ethers in Settled House Dust From Urban Dwellings With Resident Preschool-Aged Children in Nanjing, China. Archives of Environmental Contamination and Toxicology 2015; 68: 9-19. DOI: 10.1007/s00244-014-0065-z. 32 58. Doherty BT, Hammel SC, Daniels JL, et al. Organophosphate Esters: Are These Flame Retardants and Plasticizers Affecting Children’s Health? Curr Environ Health Rep 2019; 6: 201-213. DOI: 10.1007/s40572-019-00258-0. 59. Pantelaki I and Voutsa D. Organophosphate flame retardants (OPFRs): A review on analytical methods and occurrence in wastewater and aquatic environment. Sci Total Environ 2019; 649: 247-263. 2018/09/03. DOI: 10.1016/j.scitotenv.2018.08.286. 60. Du J, Li H, Xu S, et al. A review of organophosphorus flame retardants (OPFRs): occurrence, bioaccumulation, toxicity, and organism exposure. Environ Sci Pollut Res Int 2019; 26: 22126-22136. 2019/06/28. DOI: 10.1007/s11356-019-05669-y. 61. Wang Y, Li W, Martínez-Moral MP, et al. Metabolites of organophosphate esters in urine from the United States: Concentrations, temporal variability, and exposure assessment. Environ Int 2019; 122: 213-221. 2018/11/20. DOI: 10.1016/j.envint.2018.11.007. 62. Stapleton HM, Klosterhaus S, Keller A, et al. Identification of flame retardants in polyurethane foam collected from baby products. Environ Sci Technol 2011; 45: 5323-5331. 2011/05/20. DOI: 10.1021/es2007462. 63. Sundkvist AM, Olofsson U and Haglund P. Organophosphorus flame retardants and plasticizers in marine and fresh water biota and in human milk. J Environ Monit 2010; 12: 943-951. 2010/04/13. DOI: 10.1039/b921910b. 64. Wei GL, Li DQ, Zhuo MN, et al. Organophosphorus flame retardants and plasticizers: sources, occurrence, toxicity and human exposure. Environ Pollut 2015; 196: 29-46. 2014/10/08. DOI: 10.1016/j.envpol.2014.09.012. 65. Bollmann UE, Möller A, Xie Z, et al. Occurrence and fate of organophosphorus flame retardants and plasticizers in coastal and marine surface waters. Water research 2012; 46: 531-538. 66. Dodson RE, Rodgers KM, Carey G, et al. Flame retardant chemicals in college dormitories: flammability standards influence dust concentrations. Environmental science & technology 2017; 51: 4860-4869. 67. Giulivo M, Capri E, Kalogianni E, et al. Occurrence of halogenated and organophosphate flame retardants in sediment and fish samples from three European river basins. Science of the Total Environment 2017; 586: 782-791. 68. Ali N, Malik RN, Mehdi T, et al. Organohalogenated contaminants (OHCs) in the serum and hair of pet cats and dogs: biosentinels of indoor pollution. Science of the total environment 2013; 449: 29-36. 69. Yang F, Ding J, Huang W, et al. Particle size-specific distributions and preliminary exposure assessments of organophosphate flame retardants in office air particulate matter. Environmental science & technology 2014; 48: 63-70. 70. Mö ller A, Sturm R, Xie Z, et al. Organophosphorus flame retardants and plasticizers in airborne particles over the Northern Pacific and Indian Ocean toward the polar regions: Evidence for global occurrence. Environmental science & technology 2012; 46: 3127-3134. 71. Kim J-W, Isobe T, Chang K-H, et al. Levels and distribution of organophosphorus flame retardants and plasticizers in fishes from Manila Bay, the Philippines. Environmental pollution 2011; 159: 3653-3659. 72. Bika SH, Adeniji AO, Okoh AI, et al. Spatiotemporal Distribution and Analysis of Organophosphate Flame Retardants in the Environmental Systems: A Review. Molecules 2022; 27: 573. DOI: 10.3390/molecules27020573. 73. Li W, Wang Y, Asimakopoulos AG, et al. Organophosphate esters in indoor dust from 12 countries: Concentrations, composition profiles, and human exposure. Environ Int 2019; 133: 105178. DOI: https://doi.org/10.1016/j.envint.2019.105178. 74. Castorina R, Butt C, Stapleton HM, et al. Flame retardants and their metabolites in the homes and urine of pregnant women residing in California (the CHAMACOS cohort). Chemosphere 2017; 179: 159- 166. 2017/03/22. DOI: 10.1016/j.chemosphere.2017.03.076. 33 75. Cequier E, Sakhi AK, Marcé RM, et al. Human exposure pathways to organophosphate triesters - a biomonitoring study of mother-child pairs. Environ Int 2015; 75: 159-165. 2014/12/03. DOI: 10.1016/j.envint.2014.11.009. 76. Hoffman K, Daniels JL and Stapleton HM. Urinary metabolites of organophosphate flame retardants and their variability in pregnant women. Environ Int 2014; 63: 169-172. 2013/12/10. DOI: 10.1016/j.envint.2013.11.013. 77. Ospina M, Jayatilaka NK, Wong LY, et al. Exposure to organophosphate flame retardant chemicals in the U.S. general population: Data from the 2013-2014 National Health and Nutrition Examination Survey. Environ Int 2018; 110: 32-41. 2017/11/06. DOI: 10.1016/j.envint.2017.10.001. 78. Hoffman K, Lorenzo A, Butt CM, et al. Predictors of urinary flame retardant concentration among pregnant women. Environ Int 2017; 98: 96-101. 2016/10/13. DOI: 10.1016/j.envint.2016.10.007. 79. Ingle ME, Watkins D, Rosario Z, et al. The association of urinary organophosphate ester metabolites and self-reported personal care and household product use among pregnant women in Puerto Rico. Environ Res 2019; 179: 108756. 2019/10/02. DOI: 10.1016/j.envres.2019.108756. 80. Percy Z, Vuong AM, Ospina M, et al. Organophosphate esters in a cohort of pregnant women: variability and predictors of exposure. Environmental research 2020; 184: 109255. 81. Romano ME, Hawley NL, Eliot M, et al. Variability and predictors of urinary concentrations of organophosphate flame retardant metabolites among pregnant women in Rhode Island. Environ Health 2017; 16: 40-40. DOI: 10.1186/s12940-017-0247-z. 82. Romano ME, Hawley NL, Eliot M, et al. Variability and predictors of urinary concentrations of organophosphate flame retardant metabolites among pregnant women in Rhode Island. Environmental Health 2017; 16: 1-11. 83. Varshavsky JR, Robinson JF, Zhou Y, et al. Organophosphate Flame Retardants, Highly Fluorinated Chemicals, and Biomarkers of Placental Development and Disease During Mid-Gestation. Toxicol Sci 2021; 181: 215-228. 2021/03/08. DOI: 10.1093/toxsci/kfab028. 84. Ding J, Xu Z, Huang W, et al. Organophosphate ester flame retardants and plasticizers in human placenta in Eastern China. Sci Total Environ 2016; 554-555: 211-217. 2016/03/08. DOI: 10.1016/j.scitotenv.2016.02.171. 85. Zhao F, Chen M, Gao F, et al. Organophosphorus Flame Retardants in Pregnant Women and Their Transfer to Chorionic Villi. Environ Sci Technol 2017; 51: 6489-6497. 2017/05/19. DOI: 10.1021/acs.est.7b01122. 86. Bai XY, Lu SY, Xie L, et al. A pilot study of metabolites of organophosphorus flame retardants in paired maternal urine and amniotic fluid samples: potential exposure risks of tributyl phosphate to pregnant women. Environ Sci Process Impacts 2019; 21: 124-132. 2018/11/15. DOI: 10.1039/c8em00389k. 87. Patisaul HB, Behl M, Birnbaum LS, et al. Beyond Cholinesterase Inhibition: Developmental Neurotoxicity of Organophosphate Ester Flame Retardants and Plasticizers. Environ Health Perspect 2021; 129: 105001. 2021/10/07. DOI: 10.1289/ehp9285. 88. Colborn T. Neurodevelopment and endocrine disruption. Environ Health Perspect 2004; 112: 944-949. 2004/06/17. DOI: 10.1289/ehp.6601. 89. Umezu T, Yonemoto J, Soma Y, et al. Tris (2-chloroethyl) phosphate increases ambulatory activity in mice: pharmacological analyses of its neurochemical mechanism. Toxicology and applied pharmacology 1998; 148: 109-116. 90. Hausherr V, van Thriel C, Krug A, et al. Impairment of glutamate signaling in mouse central nervous system neurons in vitro by tri-ortho-cresyl phosphate at noncytotoxic concentrations. Toxicological Sciences 2014; 142: 274-284. 91. Hogberg HT, de Cássia da Silveira E Sá R, Kleensang A, et al. Organophosphorus flame retardants are developmental neurotoxicants in a rat primary brainsphere in vitro model. Archives of Toxicology 2021; 95: 207-228. DOI: 10.1007/s00204-020-02903-2. 92. Sheffler ZM, Reddy V and Pillarisetty LS. Physiology, Neurotransmitters. StatPearls. Treasure Island (FL): StatPearls Publishing 34 Copyright © 2023, StatPearls Publishing LLC., 2023. 93. Yang W, Zhao F, Fang Y, et al. 1H-nuclear magnetic resonance metabolomics revealing the intrinsic relationships between neurochemical alterations and neurobehavioral and neuropathological abnormalities in rats exposed to tris(2-chloroethyl)phosphate. Chemosphere 2018; 200: 649-659. DOI: https://doi.org/10.1016/j.chemosphere.2018.02.056. 94. Liu X, Zhao X, Wang Y, et al. Triphenyl phosphate permeates the blood brain barrier and induces neurotoxicity in mouse brain. Chemosphere 2020; 252: 126470. DOI: https://doi.org/10.1016/j.chemosphere.2020.126470. 95. Rock KD and Patisaul HB. Environmental Mechanisms of Neurodevelopmental Toxicity. Curr Environ Health Rep 2018; 5: 145-157. DOI: 10.1007/s40572-018-0185-0. 96. Rock KD, St Armour G, Horman B, et al. Effects of Prenatal Exposure to a Mixture of Organophosphate Flame Retardants on Placental Gene Expression and Serotonergic Innervation in the Fetal Rat Brain. Toxicol Sci 2020; 176: 203-223. 2020/04/04. DOI: 10.1093/toxsci/kfaa046. 97. Behl M, Hsieh J-H, Shafer TJ, et al. Use of alternative assays to identify and prioritize organophosphorus flame retardants for potential developmental and neurotoxicity. Neurotoxicology and teratology 2015; 52: 181-193. 98. Zhang S, Ireland D, Sipes NS, et al. Screening for neurotoxic potential of 15 flame retardants using freshwater planarians. Neurotoxicology and Teratology 2019; 73: 54-66. 99. Glazer L, Wells CN, Drastal M, et al. Developmental exposure to low concentrations of two brominated flame retardants, BDE-47 and BDE-99, causes life-long behavioral alterations in zebrafish. Neurotoxicology 2018; 66: 221-232. 100. Glazer L, Hawkey AB, Wells CN, et al. Developmental exposure to low concentrations of organophosphate flame retardants causes life-long behavioral alterations in zebrafish. Toxicological Sciences 2018; 165: 487-498. 101. Berger M, Gray JA and Roth BL. The expanded biology of serotonin. Annu Rev Med 2009; 60: 355-366. 2009/07/28. DOI: 10.1146/annurev.med.60.042307.110802. 102. Rangel-Barajas C, Coronel I and Florán B. Dopamine Receptors and Neurodegeneration. Aging Dis 2015; 6: 349-368. 2015/10/02. DOI: 10.14336/ad.2015.0330. 103. Sun L, Xu W, Peng T, et al. Developmental exposure of zebrafish larvae to organophosphate flame retardants causes neurotoxicity. Neurotoxicol Teratol 2016; 55: 16-22. 2016/03/29. DOI: 10.1016/j.ntt.2016.03.003. 104. Dishaw LV, Powers CM, Ryde IT, et al. Is the PentaBDE replacement, tris (1,3-dichloro-2- propyl) phosphate (TDCPP), a developmental neurotoxicant? Studies in PC12 cells. Toxicology and applied pharmacology 2011; 256: 281-289. 2011/01/19. DOI: 10.1016/j.taap.2011.01.005. 105. Melymuk LE and Bláha L. Prioritization of hazards of novel flame retardants using the mechanistic toxicology information from ToxCast and Adverse Outcome Pathways. 2019. 106. Zhong X, Wu J, Ke W, et al. Neonatal exposure to organophosphorus flame retardant TDCPP elicits neurotoxicity in mouse hippocampus via microglia-mediated inflammation in vivo and in vitro. Arch Toxicol 2020; 94: 541-552. 2020/01/03. DOI: 10.1007/s00204-019-02635-y. 107. Program NT. NTP Toxicology and Carcinogenesis Studies of Tris (2-chloroethyl) Phosphate (CAS No. 115-96-8) in F344/N Rats and B6C3F1 Mice (Gavage Studies). National Toxicology Program technical report series 1991; 391: 1-233. 108. Tilson H, Veronesi B, McLamb R, et al. Acute exposure to tris (2-chloroethyl) phosphate produces hippocampal neuronal loss and impairs learning in rats. Toxicology and applied pharmacology 1990; 106: 254-269. 109. Zhong X, Yu Y, Wang C, et al. Hippocampal proteomic analysis reveals the disturbance of synaptogenesis and neurotransmission induced by developmental exposure to organophosphate flame retardant triphenyl phosphate. J Hazard Mater 2021; 404: 124111. 2020/11/15. DOI: 10.1016/j.jhazmat.2020.124111. 35 110. Chen H, Wang P, Du Z, et al. Oxidative stress, cell cycle arrest, DNA damage and apoptosis in adult zebrafish (Danio rerio) induced by tris(1,3-dichloro-2-propyl) phosphate. Aquat Toxicol 2018; 194: 37-45. 2017/11/18. DOI: 10.1016/j.aquatox.2017.11.001. 111. Yan S, Wu H, Qin J, et al. Halogen-free organophosphorus flame retardants caused oxidative stress and multixenobiotic resistance in Asian freshwater clams (Corbicula fluminea). Environ Pollut 2017; 225: 559-568. 2017/03/21. DOI: 10.1016/j.envpol.2017.02.071. 112. Ross BM, McKenzie I, Glen I, et al. Increased levels of ethane, a non-invasive marker of n-3 fatty acid oxidation, in breath of children with attention deficit hyperactivity disorder. Nutritional neuroscience 2003; 6: 277-281. 113. Carter C. Schizophrenia susceptibility genes converge on interlinked pathways related to glutamatergic transmission and long-term potentiation, oxidative stress and oligodendrocyte viability. Schizophrenia research 2006; 86: 1-14. 114. James S. Increased oxidative stress and impaired methylation capacity in children with autism: metabolic biomarkers and genetic predisposition. Fall DAN 2004: 143-159. 115. Wood SJ, Yücel M, Pantelis C, et al. Neurobiology of schizophrenia spectrum disorders: the role of oxidative stress. Ann Acad Med Singapore 2009; 38: 396-401. 116. Taurines R, Thome J, Duvigneau JC, et al. Expression analyses of the mitochondrial complex I 75-kDa subunit in early onset schizophrenia and autism spectrum disorder: increased levels as a potential biomarker for early onset schizophrenia. European child & adolescent psychiatry 2010; 19: 441-448. 117. di Michele F, Prichep L, John ER, et al. The neurophysiology of attention-deficit/hyperactivity disorder. International Journal of Psychophysiology 2005; 58: 81-93. 118. Patisaul HB, Roberts SC, Mabrey N, et al. Accumulation and endocrine disrupting effects of the flame retardant mixture Firemaster® 550 in rats: an exploratory assessment. J Biochem Mol Toxicol 2013; 27: 124-136. 2012/11/10. DOI: 10.1002/jbt.21439. 119. Rock KD, Horman B, Phillips AL, et al. EDC IMPACT: Molecular effects of developmental FM 550 exposure in Wistar rat placenta and fetal forebrain. Endocrine connections 2018; 7: 305. 120. Liu X, Ji K and Choi K. Endocrine disruption potentials of organophosphate flame retardants and related mechanisms in H295R and MVLN cell lines and in zebrafish. Aquatic toxicology 2012; 114: 173- 181. 121. Dishaw LV, J Macaulay L, Roberts SC, et al. Exposures, mechanisms, and impacts of endocrine- active flame retardants. Current Opinion in Pharmacology 2014; 19: 125-133. DOI: https://doi.org/10.1016/j.coph.2014.09.018. 122. Farhat A, Crump D, Chiu S, et al. In ovo effects of two organophosphate flame retardants—TCPP and TDCPP—on pipping success, development, mRNA expression, and thyroid hormone levels in chicken embryos. toxicological sciences 2013; 134: 92-102. 123. Preston EV, McClean MD, Claus Henn B, et al. Associations between urinary diphenyl phosphate and thyroid function. Environ Int 2017; 101: 158-164. 2017/02/07. DOI: 10.1016/j.envint.2017.01.020. 124. Liu X, Ji K, Jo A, et al. Effects of TDCPP or TPP on gene transcriptions and hormones of HPG axis, and their consequences on reproduction in adult zebrafish (Danio rerio). Aquat Toxicol 2013; 134- 135: 104-111. 2013/04/23. DOI: 10.1016/j.aquatox.2013.03.013. 125. Liu X, Jung D, Jo A, et al. Long-term exposure to triphenylphosphate alters hormone balance and HPG, HPI, and HPT gene expression in zebrafish (Danio rerio). Environ Toxicol Chem 2016; 35: 2288- 2296. 2016/02/13. DOI: 10.1002/etc.3395. 126. Tao Y, Hu L, Liu L, et al. Prenatal exposure to organophosphate esters and neonatal thyroid- stimulating hormone levels: A birth cohort study in Wuhan, China. Environ Int 2021; 156: 106640. 2021/05/21. DOI: 10.1016/j.envint.2021.106640. 127. Zuñiga LFF, Muñoz YS and Pustovrh MC. Thyroid hormones: Metabolism and transportation in the fetoplacental unit. Molecular Reproduction and Development 2022; 89: 526-539. DOI: https://doi.org/10.1002/mrd.23647. 36 128. Eng L and Lam L. Thyroid Function During the Fetal and Neonatal Periods. Neoreviews 2020; 21: e30-e36. 2020/01/03. DOI: 10.1542/neo.21-1-e30. 129. Korevaar TIM, Muetzel R, Medici M, et al. Association of maternal thyroid function during early pregnancy with offspring IQ and brain morphology in childhood: a population-based prospective cohort study. The Lancet Diabetes & Endocrinology 2016; 4: 35-43. DOI: https://doi.org/10.1016/S2213- 8587(15)00327-7. 130. Sun M, Cao X, Wu Y, et al. Prenatal exposure to endocrine-disrupting chemicals and thyroid function in neonates: A systematic review and meta-analysis. Ecotoxicology and Environmental Safety 2022; 231: 113215. DOI: https://doi.org/10.1016/j.ecoenv.2022.113215. 131. Baldwin KR, Phillips AL, Horman B, et al. Sex Specific Placental Accumulation and Behavioral Effects of Developmental Firemaster 550 Exposure in Wistar Rats. Sci Rep 2017; 7: 7118. 2017/08/05. DOI: 10.1038/s41598-017-07216-6. 132. Leonetti C, Butt CM, Hoffman K, et al. Brominated flame retardants in placental tissues: associations with infant sex and thyroid hormone endpoints. Environmental Health 2016; 15: 113. DOI: 10.1186/s12940-016-0199-8. 133. Choi G, Keil AP, Richardson DB, et al. Pregnancy exposure to organophosphate esters and the risk of attention-deficit hyperactivity disorder in the Norwegian mother, father and child cohort study. Environ Int 2021; 154: 106549. DOI: https://doi.org/10.1016/j.envint.2021.106549. 134. Prezioso G, Giannini C and Chiarelli F. Effect of Thyroid Hormones on Neurons and Neurodevelopment. Horm Res Paediatr 2018; 90: 73-81. 2018/08/30. DOI: 10.1159/000492129. 135. Luu TM, Rehman Mian MO and Nuyt AM. Long-Term Impact of Preterm Birth: Neurodevelopmental and Physical Health Outcomes. Clin Perinatol 2017; 44: 305-314. 2017/05/10. DOI: 10.1016/j.clp.2017.01.003. 136. Pinto‐Martin J, Whitaker A, Feldman J, et al. Special education services and school performance in a regional cohort of low‐birthweight infants at age nine. Paediatric and perinatal epidemiology 2004; 18: 120-129. 137. Hua J, Barnett AL, Lin Y, et al. Association of Gestational Age at Birth With Subsequent Neurodevelopment in Early Childhood: A National Retrospective Cohort Study in China. Front Pediatr 2022; 10: 860192. 2022/06/18. DOI: 10.3389/fped.2022.860192. 138. Rose O, Blanco E, Martinez SM, et al. Developmental Scores at 1 Year With Increasing Gestational Age, 37–41 Weeks. Pediatrics 2013; 131: e1475-e1481. DOI: 10.1542/peds.2012-3215. 139. Espel EV, Glynn LM, Sandman CA, et al. Longer gestation among children born full term influences cognitive and motor development. PLoS One 2014; 9: e113758. 140. Cortese M, Moster D and Wilcox AJ. Term Birth Weight and Neurodevelopmental Outcomes. Epidemiology 2021; 32: 583-590. 2021/05/19. DOI: 10.1097/ede.0000000000001350. 141. Basso O, Wilcox AJ and Weinberg CR. Birth weight and mortality: causality or confounding? American journal of epidemiology 2006; 164: 303-311. 142. Rosenmai AK, Winge SB, Möller M, et al. Organophosphate ester flame retardants have antiandrogenic potential and affect other endocrine related endpoints in vitro and in silico. Chemosphere 2021; 263: 127703. 2020/08/28. DOI: 10.1016/j.chemosphere.2020.127703. 143. Kojima H, Takeuchi S, Itoh T, et al. In vitro endocrine disruption potential of organophosphate flame retardants via human nuclear receptors. Toxicology 2013; 314: 76-83. 144. Ma Z, Tang S, Su G, et al. Effects of tris (2-butoxyethyl) phosphate (TBOEP) on endocrine axes during development of early life stages of zebrafish (Danio rerio). Chemosphere 2016; 144: 1920-1927. 2015/11/08. DOI: 10.1016/j.chemosphere.2015.10.049. 145. Blazer S, Moreh-Waterman Y, Miller-Lotan R, et al. Maternal hypothyroidism may affect fetal growth and neonatal thyroid function. Obstet Gynecol 2003; 102: 232-241. 2003/08/09. DOI: 10.1016/s0029-7844(03)00513-1. 146. Moore TA, Ahmad IM and Zimmerman MC. Oxidative Stress and Preterm Birth: An Integrative Review. Biol Res Nurs 2018; 20: 497-512. 2018/08/03. DOI: 10.1177/1099800418791028. 37 147. Ferguson KK, Kamai EM, Cantonwine DE, et al. Associations between repeated ultrasound measures of fetal growth and biomarkers of maternal oxidative stress and inflammation in pregnancy. Am J Reprod Immunol 2018; 80: e13017-e13017. 2018/07/08. DOI: 10.1111/aji.13017. 148. McElrath TF, Hecht JL, Dammann O, et al. Pregnancy disorders that lead to delivery before the 28th week of gestation: an epidemiologic approach to classification. American journal of epidemiology 2008; 168: 980-989. 149. Bramham K, Parnell B, Nelson-Piercy C, et al. Chronic hypertension and pregnancy outcomes: systematic review and meta-analysis. BMJ 2014; 348. 150. Davis LK, Pierce AL, Hiramatsu N, et al. Gender-specific expression of multiple estrogen receptors, growth hormone receptors, insulin-like growth factors and vitellogenins, and effects of 17β- estradiol in the male tilapia (Oreochromis mossambicus). General and Comparative Endocrinology 2008; 156: 544-551. DOI: https://doi.org/10.1016/j.ygcen.2008.03.002. 151. Guan R, Li N, Wang W, et al. The adverse outcome pathway (AOP) of estrogen interference effect induced by triphenyl phosphate (TPP): Integrated multi-omics and molecular dynamics approaches. Ecotoxicology and Environmental Safety 2022; 234: 113387. DOI: https://doi.org/10.1016/j.ecoenv.2022.113387. 152. Liu X, Cai Y, Wang Y, et al. Effects of tris(1,3-dichloro-2-propyl) phosphate (TDCPP) and triphenyl phosphate (TPP) on sex-dependent alterations of thyroid hormones in adult zebrafish. Ecotoxicol Environ Saf 2019; 170: 25-32. 2018/12/06. DOI: 10.1016/j.ecoenv.2018.11.058. 153. Sun L, Tan H, Peng T, et al. Developmental neurotoxicity of organophosphate flame retardants in early life stages of Japanese medaka (Oryzias latipes). Environ Toxicol Chem 2016; 35: 2931-2940. 2016/05/06. DOI: 10.1002/etc.3477. 154. Zhu Y, Ma X, Su G, et al. Environmentally Relevant Concentrations of the Flame Retardant Tris(1,3-dichloro-2-propyl) Phosphate Inhibit Growth of Female Zebrafish and Decrease Fecundity. Environ Sci Technol 2015; 49: 14579-14587. 2015/10/30. DOI: 10.1021/acs.est.5b03849. 155. Li H, Su G, Zou M, et al. Effects of Tris(1,3-dichloro-2-propyl) Phosphate on Growth, Reproduction, and Gene Transcription of Daphnia magna at Environmentally Relevant Concentrations. Environmental Science & Technology 2015; 49: 12975-12983. DOI: 10.1021/acs.est.5b03294. 156. Chen G, Jin Y, Wu Y, et al. Exposure of male mice to two kinds of organophosphate flame retardants (OPFRs) induced oxidative stress and endocrine disruption. Environ Toxicol Pharmacol 2015; 40: 310-318. 2015/07/18. DOI: 10.1016/j.etap.2015.06.021. 157. Yu L, Jia Y, Su G, et al. Parental transfer of tris(1,3-dichloro-2-propyl) phosphate and transgenerational inhibition of growth of zebrafish exposed to environmentally relevant concentrations. Environ Pollut 2017; 220: 196-203. 2016/09/21. DOI: 10.1016/j.envpol.2016.09.039. 158. Xu Q, Wu D, Dang Y, et al. Reproduction impairment and endocrine disruption in adult zebrafish (Danio rerio) after waterborne exposure to TBOEP. Aquat Toxicol 2017; 182: 163-171. 2016/12/03. DOI: 10.1016/j.aquatox.2016.11.019. 159. Zeng X, Sun H, Huang Y, et al. Effects of environmentally relevant concentrations of tris (2- butoxyethyl) phosphate on growth and transcription of genes involved in the GH/IGF and HPT axes in zebrafish (Danio rerio). Chemosphere 2018; 212: 376-384. 2018/08/28. DOI: 10.1016/j.chemosphere.2018.08.102. 160. Castorina R, Bradman A, Stapleton HM, et al. Current-use flame retardants: Maternal exposure and neurodevelopment in children of the CHAMACOS cohort. Chemosphere 2017; 189: 574-580. 2017/10/01. DOI: 10.1016/j.chemosphere.2017.09.037. 161. Doherty BT, Hoffman K, Keil AP, et al. Prenatal exposure to organophosphate esters and cognitive development in young children in the Pregnancy, Infection, and Nutrition Study. Environ Res 2019; 169: 33-40. 2018/11/09. DOI: 10.1016/j.envres.2018.10.033. 162. Doherty BT, Hoffman K, Keil AP, et al. Prenatal exposure to organophosphate esters and behavioral development in young children in the Pregnancy, Infection, and Nutrition Study. Neurotoxicology 2019; 73: 150-160. 2019/04/06. DOI: 10.1016/j.neuro.2019.03.007. 38 163. Liu W, Luo D, Xia W, et al. Prenatal exposure to halogenated, aryl, and alkyl organophosphate esters and child neurodevelopment at two years of age. J Hazard Mater 2020; 408: 124856. 2021/01/01. DOI: 10.1016/j.jhazmat.2020.124856. 164. Percy Z, Chen A, Yang W, et al. Childhood urinary organophosphate esters and cognitive abilities in a longitudinal cohort study. Environmental Research 2022; 215: 114265. DOI: https://doi.org/10.1016/j.envres.2022.114265. 165. Ramos AM, Herring AH, Villanger GD, et al. The association of prenatal phthalates, organophosphorous pesticides, and organophosphate esters with early child language ability in Norway. Environmental Research 2023; 225: 115508. DOI: https://doi.org/10.1016/j.envres.2023.115508. 166. Hall AM, Ramos AM, Drover SSM, et al. Gestational organophosphate ester exposure and preschool attention-deficit/hyperactivity disorder in the Norwegian Mother, Father, and Child cohort study. International Journal of Hygiene and Environmental Health 2023; 248: 114078. DOI: https://doi.org/10.1016/j.ijheh.2022.114078. 167. Sugeng EJ, de Cock M, Leonards PEG, et al. Association of exposure to organophosphate flame retardants and children's behavior at a median age of 18 months. Environmental Advances 2021; 5: 100077. DOI: https://doi.org/10.1016/j.envadv.2021.100077. 168. Lipscomb ST, McClelland MM, MacDonald M, et al. Cross-sectional study of social behaviors in preschool children and exposure to flame retardants. Environ Health 2017; 16: 23. 2017/03/10. DOI: 10.1186/s12940-017-0224-6. 169. Hutter H-P, Haluza D, Piegler K, et al. Semivolatile compounds in schools and their influence on cognitive performance of children. International Journal of Occupational Medicine and Environmental Health 2013; 26: 628-635. journal article. DOI: 10.2478/s13382-013-0125-z. 170. Percy Z, Chen A, Sucharew H, et al. Early-life exposure to a mixture of organophosphate esters and child behavior. International Journal of Hygiene and Environmental Health 2023; 250: 114162. DOI: https://doi.org/10.1016/j.ijheh.2023.114162. 171. Percy Z, Vuong AM, Xu Y, et al. Prenatal exposure to a mixture of organophosphate esters and intelligence among 8-year-old children of the HOME Study. NeuroToxicology 2021; 87: 149-155. DOI: https://doi.org/10.1016/j.neuro.2021.09.005. 172. Gan H, Zhang Y, Wang Y-f, et al. Relationships of prenatal organophosphate ester exposure with pregnancy and birth outcomes: A systematic scoping review of epidemiological studies. Ecotoxicology and Environmental Safety 2023; 252: 114642. DOI: https://doi.org/10.1016/j.ecoenv.2023.114642. 173. Hoffman K, Stapleton HM, Lorenzo A, et al. Prenatal exposure to organophosphates and associations with birthweight and gestational length. Environ Int 2018; 116: 248-254. 2018/04/27. DOI: 10.1016/j.envint.2018.04.016. 174. Luo D, Liu W, Tao Y, et al. Prenatal Exposure to Organophosphate Flame Retardants and the Risk of Low Birth Weight: A Nested Case-Control Study in China. Environ Sci Technol 2020; 54: 3375- 3385. 2020/02/29. DOI: 10.1021/acs.est.9b06026. 175. Luo D, Liu W, Wu W, et al. Trimester-specific effects of maternal exposure to organophosphate flame retardants on offspring size at birth: A prospective cohort study in China. Journal of Hazardous Materials 2021; 406: 124754. DOI: https://doi.org/10.1016/j.jhazmat.2020.124754. 176. Crawford KA, Hawley N, Calafat AM, et al. Maternal urinary concentrations of organophosphate ester metabolites: associations with gestational weight gain, early life anthropometry, and infant eating behaviors among mothers-infant pairs in Rhode Island. Environmental Health 2020; 19: 97. DOI: 10.1186/s12940-020-00648-0. 177. Kuiper JR, Stapleton HM, Wills-Karp M, et al. Predictors and reproducibility of urinary organophosphate ester metabolite concentrations during pregnancy and associations with birth outcomes in an urban population. Environ Health 2020; 19: 55. 2020/05/26. DOI: 10.1186/s12940-020-00610-0. 178. Bommarito PA, Welch BM, Keil AP, et al. Prenatal exposure to consumer product chemical mixtures and size for gestational age at delivery. Environ Health 2021; 20: 68-68. DOI: 10.1186/s12940- 021-00724-z. 39 179. Yang W, Braun JM, Vuong AM, et al. Gestational exposure to organophosphate esters and infant anthropometric measures in the first 4 weeks after birth. Science of The Total Environment 2023; 857: 159322. DOI: https://doi.org/10.1016/j.scitotenv.2022.159322. 40 CHAPTER 2 Methods Overview MADRES Cohort Description All maternal-infant dyads analyzed in the three studies of this dissertation are participants from the Maternal and Developmental Risks from Environmental and Social Stressors (MADRES) pregnancy cohort. The MADRES cohort study is an ongoing prospective study of approximately 1000 predominately low-income, Hispanic women and children living in urban Los Angeles, California. A detailed description of the MADRES study population and protocol have been described elsewhere. 1 In brief, MADRES recruitment began in November 2015 and participants have been recruited from two non-profit community health clinics, one county hospital prenatal clinic, one private obstetrics and gynecology practice, and through self-referral from community meetings and local advertisements. Eligible participants are less than 30 weeks’ gestation at the time of recruitment, over 18 years of age, and fluent English or Spanish speakers. Exclusion criteria for study participation includes (1) multiple gestation, (2) having a physical, mental, or cognitive disability that would prevent participation or ability to provide consent, (3) current incarceration, or (4) HIV positive status. By August 2022, a total of 774 maternal participants had delivered their infants in the MADRES cohort. Maternal and infant demographics are shown in Table 1. Table 1: MADRES Cohort Participant Characteristics (N=774) Mean (SD)/Freq(%) Maternal Characteristics Recruitment Site LAC+USC Eisner USC Obstetrics & Gynecology Community Recruit South Central Clinic 151 (19.5%) 536 (69.3%) 58 (7.5%) 5 (0.65%) 24 (3.1%) Cohort <20 weeks (Regular Entry) 20+ weeks (Late Entry) 553 (71.5%) 221 (28.6%) Age 28.3 (6.0) Nativity Non-Hispanic US-Born Hispanic Foreign-Born Hispanic Missing 153 (19.8%) 238 (30.7%) 274 (35.4%) 109 (14.1%) 41 Education Less than 12 th grade Completed grade 12 Some college or technical school Completed 4 years of college Some graduate training after college Missing 186 (24.0%) 230 (29.7%) 198 (25.6%) 78 (10.1%) 40 (5.2%) 42 (5.4%) Income Don’t Know Less than $15,000 $15,000 to $29,999 $30,000 to $49,999 $50,000 to $99,999 $100,000+ Missing 244 (31.5%) 157 (20.3%) 177 (22.9%) 82 (10.6%) 38 (4.9%) 35 (4.5%) 41 (5.3%) Language English Spanish Missing 519 (67.1%) 253 (32.7%) 2 (0.3%) Hispanic Ethnicity No Yes Missing 153 (19.8%) 579 (74.8%) 42 (5.4%) Race×Ethnicity White, non-Hispanic Black, non-Hispanic Hispanic Multiracial, non-Hispanic Other, non-Hispanic Missing 40 (5.2%) 87 (11.2%) 579 (74.8%) 9 (1.2%) 17 (2.2%) 42 (5.4%) Infant Characteristics Sex Female Male Missing 384 (49.6%) 387 (50.0%) 3 (0.4%) Delivery Method Vaginal Delivery Planned C-Section Unplanned/Emergency C-Section Vaginal Birth After Cesarean Vacuum Assisted Vaginal Forceps Assisted Vaginal Missing 498 (64.3%) 100 (12.9%) 105 (13.6%) 39 (5.0%) 18 (2.3%) 1 (0.1%) 13 (1.7%) Gestational Age at Birth 39.0 (1.8) Birthweight 3267.0 (520.5) 42 Relevant Methodology MADRES data are collected via interviewer-administered questionnaires, anthropometric measurements, and biospecimen samples during in person visits and telephone questionnaires and through medical record abstractions. Participants are followed through their pregnancies, at birth, and through their child’s first 5 years of life. Data collected throughout the various stages of this cohort were used in the three studies of this dissertation, including data on prenatal concentrations of OPE metabolites, birth outcomes, motor development during various timepoints in infancy, and child neurobehavioral development. Prenatal Measures of Urinary OPE Metabolites Prenatal exposures to OPEs in 426 maternal participants were measured in spot urine samples collected in 90 mL sterile specimen containers during the third trimester pregnancy timepoint. Urine specimens were separated into 1.5 mL aliquot cryovials and stored at -80 ºCelsius prior to shipment. Specific gravity was measured in urine samples using a digital handheld refractometer (ATAGO PAL-10s pocket refractometer). Urinary samples were sent to the Wadsworth Center Human Health Exposure Analysis Resource (HHEAR) for the analysis of the following nine OPE metabolites: diphenyl phosphate (DPHP), sum of dibutyl phosphate (DNBP) and di-isobutyl phosphate (DIBP), bis(1,3,-dichloro-2-propyl) phosphate (BDCIPP), bis(2-chloroethyl) phosphate (BCEP), bis(butoxethyl) phosphate (BBOEP), bis(1- chloro-2-propyl) phosphate (BCIPP), bis(2-ethylhexyl) phosphate (BEHP), bis(2-methylphenyl) phosphate (BMPP), dipropyl phosphate (DPRP). The nine OPE metabolites were analyzed following methods similar to those described elsewhere with some modifications. 2 Briefly, urine samples (0.5 mL) were aliquoted into pre-baked glass tubes and spiked with 1 ng of deuterated internal standard (IS) mixtures of OPEs and 1 mL of 10 mM ammonium acetate buffer (pH 5). The samples were passed through solid phase extraction (SPE) cartridges (STRATA- X-AW: 60 mg, 3cc, Phenomenex, Torrance, CA, USA) which were conditioned by successive passage with 2 mL of 5% (v/v) ammonia/methanol, 2 mL of methanol, and 2 mL of water. The samples were loaded with the valves partially opened. The SPE cartridges were then dried under vacuum for 3 min after washing with 43 1.0 mL of water. Analytes were eluted with 2 times 0.5 mL of 5% (v/v) ammonia/methanol, concentrated under a gentle stream of nitrogen at 37 °C to near dryness, and reconstituted with 0.1 mL of acetonitrile. High-performance liquid chromatography (HPLC, ExionLC™ system; SCIEX, Redwood City, CA, USA), coupled with an AB SCIEX QTRAP 5500+ triple quadrupole mass spectrometer (Applied Biosystems, Foster City, CA, USA), was used in the identification and quantification of target compounds. Nine OPE diester metabolites and corresponding 9 internal standards were separated by a Kinetex HILIC column (100 mm × 2.1 mm, 2.6 μm particle size; Phenomenex) serially connected to a Betasil C18 guard column (20 mm × 2.1 mm, 5 μm particle size; Thermo Scientific). The analytes were quantified by isotopic dilution method and an 11-point calibration curve (at concentrations ranging from 0.02 to 50 ng/mL) with the regression coefficient ≥ 0.998. Matrix spikes (synthetic and urine pool spiked with 1 ng of native standards and 1 ng of internal standards) were analyzed with real samples as quality control (QC) samples. For each batch of samples, replicates of reagent blanks, matrix blanks, and matrix spiked samples were processed. Replicates of HHEAR Urine Quality Control (QC) Pools Standard Reference Materials (SRM3672 and SRM3673, NIST, Gaithersburg, MD, USA) were analyzed with every batch of samples. Trace levels of all OPE diester metabolites were found in procedural blanks. OPE diester metabolite concentrations in samples were subtracted from their respective blank values. Matrix spiked samples had average recoveries of 70.4-133% (CV: ±9-19%). Repeated analysis of HHEAR Urine QC Pools A and B among batches showed coefficients of variation of ±12-31% and ±12-30% respectively. SRM3672 and SRM3673 had coefficients of variation of ±12-40% and ±12-27% respectively. The limit of detection (LOD) of target analytes ranged from 0.012 to 0.0441 ng/mL. Due to the poor chromatographic separation and co-elution of peaks accompanying a similar mass transition for DNBP and DIBP in our mass spectrometer, a sum concentration for the metabolites of dibutyl phosphate and di-isobutyl phosphate (DNBP + DIBP) was analyzed in this study. OPE metabolites with levels below the LOD were imputed using the LOD/√2. All metabolites were then specific gravity (SG) adjusted using the following formula: Pc=P[(SGm-1)/(SG-1)], where Pc is the specific gravity corrected toxicant concentration (ng/mL), P is the observed toxicant concentration 44 (ng/mL), SGm is the median SG value among the study population (median= 1.016), and SG= the SG value of the participant. 3 Measures of Gestational Age at Birth and Birthweight-for-GA Z scores Infant GA at birth is calculated from a variety of sources and assigned for each infant using a hierarchy of methods outlined by the American College of Obstetricians and Gynecologists. 4 Based on data availability and order of preference, the following hierarchy was used to assign infant GA at birth: 1) first trimester (<14 weeks gestation) ultrasound measurement of crown-rump-length, (2) second trimester (<28 weeks gestation) ultrasound measurement of fetal biparietal diameter, (3) physician's best clinical estimate from electronic medical records, and (4) estimated from last menstrual period. Infant birthweight was directly abstracted from medical records and sex-specific BW for GA z-scores were calculated using a nationally representative US sample. 5 Gross and Fine Motor Scores During Infancy Infant gross and fine motor development was assessed through telephone questionnaire in the participant’s preferred language at the 6-, 9-, 12-, and 18-month study timepoint using the gross and fine motor subscales of the Ages and Stages Questionnaire, Third Edition (ASQ-3). ASQ-3 administration at the four study timepoints occurred when the child turned the age specified at the study visit timepoint (6, 9, 12, 18 months) and continued until the close of that study visit window, with the following ranges in windows by study visit timepoints: (1) 6 months + 6 weeks, (2) 9 months + 6 weeks, (3) 11 months + 8 weeks, and (4) 18 months + 6 weeks, respectively. Scores for the ASQ-3 range from 0 to 60, with predetermined cutoffs defined as scores that are at least 1 standard deviation lower than the average score of a normative US sample (monitoring zone) and scores that are at least 2 standard deviations lower than the average score of a normative US sample (further reassessment with a healthcare professional). Both motor ASQ-3 subscales were administered and scored according to the ASQ-3 protocol. 6 45 Child Behavior Checklist (CBCL) Children’s neurobehavioral symptoms were measured using the Child Behavior Checklist (CBCL 1.5-5) administered at the 36-month study visit. The CBCL is a 99-item parent-reported questionnaire designed to assess the frequency of behaviors observed in their child in the past 2 months on a Likert scale (not true=0, sometimes true=1, or very often true=2), with higher scores indicating increasing problems. Syndrome scales (emotionally reactive, anxious/depressed, somatic complaints, withdrawn, sleep problems, attention problems, aggressive behaviors, and other problems) are created by summing together relevant items, and t-scores for borderline (t-scores: 60 – 63) and clinical categories (t-scores: ≥64) are calculated based on previously described criteria. 7 Syndrome scales raw scores can be summed together to create composite CBCL scales, including internalizing problems (emotionally reactive, anxious/depressed, somatic complaints, and withdrawn) and externalizing problems (attention problems and aggressive behavior), along with a total problems scale comprised of the summed total of all 99 questionnaire items, plus the highest score on any additional problems in the open-ended questions (score range=0-200). Covariates Relevant covariates of interest related to pregnancy, including maternal race/ethnicity, age at enrollment, household income, maternal smoking during pregnancy status, and education, were collected through interviewer administered questionnaires in the participant’s preferred language (Spanish or English). Maternal hypertensive disorders of pregnancy, gestational diabetes, and delivery method were abstracted from maternal electronic health records. Maternal pre-pregnancy BMI was calculated using participant reported weight and standing height was measured with a commercial stadiometer (Perspectives Enterprises model P-AIM-101) by study staff in the first study visit. Infant sex was primarily abstracted from electronic medical records but if unavailable, was maternal reported. Sample collection variables, such as gestational age at sample collection and season of sample collection, were collected at the time of urine sample collection. 46 MADRES Funding This work was supported by the Maternal and Developmental Risks from Environmental and Social Stressors (MADRES) Center (grant #s P50ES026086, 83615801, P50MD015705) funded by the National Institute of Environmental Health Sciences, the National Institute for Minority Health and Health Disparities, and the Environmental Protection Agency; the Southern California Environmental Health Sciences Center (grant # 5P30ES007048) funded by the National Institute of Environmental Health Sciences, and the Life course Approach to Developmental Repercussions of Environmental Agents on Metabolic and Respiratory health (LA DREAMERs) (grant #s UH3OD023287) funded by the National Institutes of Health Office of the Director, United States. 47 REFERENCES 1. Bastain TM, Chavez T, Habre R, et al. Study Design, Protocol and Profile of the Maternal And Developmental Risks from Environmental and Social Stressors (MADRES) Pregnancy Cohort: a Prospective Cohort Study in Predominantly Low-Income Hispanic Women in Urban Los Angeles. BMC Pregnancy Childbirth 2019; 19: 189. 2019/05/31. DOI: 10.1186/s12884-019-2330-7. 2. Wang Y, Li W, Martínez-Moral MP, et al. Metabolites of organophosphate esters in urine from the United States: Concentrations, temporal variability, and exposure assessment. Environ Int 2019; 122: 213-221. 2018/11/20. DOI: 10.1016/j.envint.2018.11.007. 3. Boeniger MF, Lowry LK and Rosenberg J. Interpretation of urine results used to assess chemical exposure with emphasis on creatinine adjustments: a review. American Industrial Hygiene Association journal 1993; 54: 615-627. 4. Dunlop AL, Essalmi AG, Alvalos L, et al. Racial and geographic variation in effects of maternal education and neighborhood-level measures of socioeconomic status on gestational age at birth: Findings from the ECHO cohorts. PloS one 2021; 16: e0245064. 5. Aris IM, Kleinman KP, Belfort MB, et al. A 2017 US reference for singleton birth weight percentiles using obstetric estimates of gestation. Pediatrics 2019; 144. 6. Squires J, Twombly,E., Bricker, D., Potter, L. The ASQ-3 Technical Report, https://agesandstages.com/resource/asq-3-technical-appendix/. (2009, 2021). 7. Achenbach TM and Rescorla LA. Manual for the ASEBA preschool forms and profiles. Burlington, VT: University of Vermont, Research center for children, youth …, 2000. 48 CHAPTER 3 STUDY 1: Sex-Specific Effects of Prenatal Organophosphate Ester (OPE) Metabolite Mixtures and Adverse Infant Birth Outcomes in the Maternal and Developmental Risks from Environmental and Social Stressors (MADRES) Pregnancy Cohort ABSTRACT Background: Organophosphate esters (OPEs) are used as flame retardants and plasticizers in various consumer products. Limited prior research suggests sex-specific effects of prenatal OPE exposures on fetal development. We evaluated overall and sex-specific associations between prenatal OPE exposures and gestational age (GA) at birth and birthweight for gestational age (BW for GA) z-scores among the predominately low-income, Hispanic MADRES cohort. Methods: Nine OPE metabolite concentrations were measured in 421 maternal urine samples collected during a third trimester visit (GA=31.5±2.0 weeks). We examined associations between single urinary OPE metabolites and GA at birth and BW for GA z-scores using linear regression models and Generalized Additive Models (GAMs) and effects from OPE mixtures using Bayesian Kernel Machine Regression (BKMR). We also assessed sex-specific differences in single metabolite analyses by evaluating statistical interactions and stratifying by sex. Results: We did not find significant associations between individual OPE metabolites and birth outcomes in the full infant sample; however, we found that higher bis(1,3-dichloro-2-propyl) phosphate (BDCIPP) was associated with earlier GA at birth among male infants (p=0.04), and a nonlinear, inverted U-shape association between the sum of dibutyl phosphate and di-isobutyl phosphate (DNBP+DIBP) and GA at birth among female infants (p=0.03). In mixtures analysis, higher OPE metabolite mixture exposures was 49 associated with lower GA at birth, which was primarily driven by female infants. No associations were observed between OPE mixtures and BW for GA z-scores. Conclusion: Higher BDCIPP and DNBP+DIBP concentrations were associated with earlier GA at birth among male and female infants, respectively. Higher exposure to OPE mixtures was associated with earlier GA at birth, particularly among female infants. However, we saw no associations between prenatal OPEs and BW for GA. Our results suggest sex-specific impacts of prenatal OPE exposures on GA at birth. INTRODUCTION Infant birth outcomes, such as gestational age (GA) at birth and birthweight, are important indicators of infant health and strong predictors of child neurodevelopment, obesity, and other cardiometabolic conditions. 1-5 Adverse birth outcomes, including preterm birth and low birthweight, disproportionately impact communities of color who are also more likely to experience a higher burden of environmental chemical exposures. 6-8 Accumulating epidemiological evidence stresses the potential for in utero environmental endocrine disrupting chemicals to adversely impact infant birth outcomes. 9-11 Emerging literature suggests that organophosphate esters (OPEs) are a class of endocrine disrupting chemicals. 12-14 OPEs have increased in use as alternative flame-retardants to replace polybrominated diphenyl ethers (PBDEs) which were phased out due to bioaccumulation and neurotoxicity concerns. 15 However, OPEs are also commonly used as plasticizers and lubricants, contributing to their environmental ubiquity. 16,17 OPEs are found in a variety of consumer, industrial, and electronic products, including clothing, polyurethane foams, textiles, and building materials. 15,18,19 Since OPEs are physically bound with a product matrix rather than chemically bound, they easily volatize and leach into surrounding environments during product use, settling into numerous environmental matrices (i.e., soil, surface water, sediment, and agricultural products) and dust particles in indoor environments, and enter the body through inhalation, dermal contact, and ingestion via dietary intake of OPE-contaminated food and drinking water. 20-26 Emerging literature indicates widespread exposure to OPEs among the general population, with high 50 detection (>95%) of metabolites such as diphenyl phosphate (DPHP; parent compound= triphenyl phosphate (TPHP)) and bis(1,3-dichloro-2-propyl) phosphate (BDCIPP; parent compound= tris(1,3- dichloropropyl) (TDCIPP)) in urine. 18,27 Widespread OPE metabolite concentrations have also recently been observed among vulnerable populations, including pregnant women. 27-32 Since OPEs have been found in the chorionic villi and uterine decidua of the placenta and amniotic fluid, suggesting in utero transfer of OPEs to the fetus, 33-36 there is a growing concern regarding exposures to these substances during critical periods of development. Animal and in-vitro toxicological studies have reported adverse reproductive outcomes and early growth and development effects, including decreased birthweight, from OPE exposures. 37-42 However, no toxicological research has been conducted on the effects of OPEs on gestational duration and preterm birth. Hypothesized biological mechanisms include endocrine disruption, oxidative stress, and placental alterations. 34,38,39,43-51 Limited epidemiological studies have explored the potential impacts of prenatal exposures to OPEs on infant birth outcomes, but the small number of studies that have been conducted have reported sex- specific effects on pre-term birth and low birthweight. 52,53 Additionally, most of the epidemiological literature on OPEs and infant birth outcomes has evaluated the effects of two OPE metabolites, DPHP and BDCIPP, in single chemical analyses, resulting in a limited understanding of the impacts of other OPEs on birth outcomes and their joint health effects. Furthermore, impacts of OPEs among structurally marginalized populations, such as Hispanic pregnant women, have been understudied. 18 This is of concern given the increasing use of OPEs and the ubiquity of exposures. 27 In this study, we evaluated the prenatal impacts of nine OPE metabolites individually and in mixtures on infant GA at birth and BW for GA among 421 mother-infant dyads participating in the Maternal And Developmental Risks from Environmental and Social Stressors (MADRES) cohort study. Given prior toxicological and epidemiological evidence suggesting adverse OPE impacts on birth outcomes, with more pronounced impacts among female infants, 38,52,53 we hypothesized that higher exposure to both single OPE 51 metabolites and mixtures would result in earlier GA at birth and lower BW for GA, particularly among female infants. METHODS Study Sample The MADRES study is an ongoing prospective pregnancy cohort of predominately low-income Hispanic/Latina women living in urban Los Angeles. A detailed description of the MADRES study population and protocol has been previously described. 54 In brief, participants were recruited into the study at three partner community health clinics, one private obstetrics and gynecology practice in Los Angeles, and through limited self-referrals from community meetings and local advertisements. Eligible participants at the time of recruitment were: (1) less than 30 weeks’ gestation, (2) over 18 years of age or older, and (3) fluent speaker of English or Spanish. Exclusion criteria included: (1) multiple gestation, (2) having a physical, mental, or cognitive disability that prevented participation or ability to provide consent, (3) current incarceration, and (4) HIV positive status. Written informed consent was obtained at study entry for each participant and the study was approved by the University of Southern California’s Institutional Review Board. Participants included in this analysis were recruited from November 2015 to October 2019. OPE metabolite concentrations were measured in all available MADRES participants’ spot urine specimens collected during the third trimester study visit (N=426, mean ± SD gestational age at collection: 31.5 weeks±2.0 weeks). Five mother-infant dyads with missing information on key covariates or birth outcomes were excluded from the final analysis. A total of 421 mother-infant dyads with OPE metabolite concentrations, birth outcomes data, and information on key covariates were analyzed in this study. This subset of 421 participants from the MADRES cohort was similar to the full cohort on key demographic characteristics. 52 OPE Metabolite Analysis Maternal single spot urine samples were collected in 90 mL sterile specimen containers. Urine specimens were aliquoted into 1.5 mL aliquot cryovials and stored at -80 ºC prior to shipment. Specific gravity was measured in urine samples prior to flash freezing using a digital handheld refractometer (ATAGO PAL-10s pocket refractometer). Frozen samples were sent to the Wadsworth Center Human Health Exposure Analysis Resource (HHEAR) for the analysis of the following nine OPE metabolites: diphenyl phosphate (DPHP), sum of dibutyl phosphate and di-isobutyl phosphate (DNBP + DIBP), bis(1,3,-dichloro-2-propyl) phosphate (BDCIPP), bis(2-chloroethyl) phosphate (BCEP), bis(butoxethyl) phosphate (BBOEP), bis(1-chloro-2-propyl) phosphate (BCIPP), bis(2-ethylhexyl) phosphate (BEHP), bis(2-methylphenyl) phosphate (BMPP), and dipropyl phosphate (DPRP). Additional information on the analyzed metabolites, their corresponding parent compounds, and common applications are described in Chapter 1 (Table 1). The nine OPE metabolites were analyzed following methods similar to those outlined in Chapter 2 (Prenatal Measures of Urinary OPE Metabolites) and described elsewhere with some modifications. 26 In brief, OPE metabolites were identified and quantified using high-performance liquid chromatography (HPLC, ExionLC™ system; SCIEX, Redwood City, CA, USA), coupled with an AB SCIEX QTRAP 5500+ triple quadrupole mass spectrometer (Applied Biosystems, Foster City, CA, USA). The limit of detection (LOD) of target analytes ranged from 0.012 to 0.0441 ng/mL. Due to the poor chromatographic separation and co-elution of peaks accompanying a similar mass transition for DNBP and DIBP in our mass spectrometer, a sum concentration for the metabolites of dibutyl phosphate and di-isobutyl phosphate (DNBP + DIBP) was analyzed in this study. OPE metabolites with levels below the LOD were imputed using the LOD/√2. 55 Metabolites were then specific gravity (SG) adjusted using the following standardization formula similarly used in prior OPE literature among other pregnancy cohorts: Pc=P[(SGm-1)/(SG-1)], where Pc is the specific gravity corrected toxicant concentration (ng/mL), P is the observed toxicant concentration (ng/mL), SGm is the 53 median SG value among the subset of participants analyzed in this study (median=1.016), and SG=the SG value of the sample. 56 Birth Outcomes Infant GA at birth was calculated using an accepted hierarchy of the following methods based on available data, in order of preference: (1) first trimester (<14 weeks gestation) ultrasound measurement of crown-rump-length (n=251, 59.6%), (2) second trimester (<28 weeks gestation) ultrasound measurement of fetal biparietal diameter (n=118, 28.0%), (3) physician’s best clinical estimate from electronic medical records (n=51, 12.1%), and (4) estimated from last menstrual period (n=1, 0.2%). 57 Many participants had multiple estimates of GA available (79%), with estimates generally highly correlated with one another across methods (Spearman 𝜌 =0.52-0.91), but higher order preference was placed on methods suggested to provide more accurate estimates of GA at birth. 57-59 Infant birthweight for all participants was directly abstracted from electronic medical records. Sex-specific BW for GA z-scores were calculated for each participant using a nationally representative U.S. sample. 60 Infant sex was primarily abstracted from electronic medical records (n=410, 97.4%), followed by maternal-reported infant sex (n=11, 2.6%) for cases in which abstracted sex could not be obtained. Covariates Potential study covariates for GA at birth and birthweight models were identified a priori based on previous literature 28,32,52,53 and visualized using Directed Acyclic Graphs (DAGs) on DAGitty (see Supplemental Figure 1a and 1b). 61 All models were adjusted for all study design and sample collection variables and variables selected in the DAG minimal sufficient adjustment sets, barring some of the exceptions discussed below. Maternal demographic and pregnancy-related covariates identified in prior literature and considered in DAGs included maternal age at study enrollment (years), income (<$30,000, ≥ $30,000, don’t know), education (≤12 th Grade, some college or technical school or completed 4 years of college, some graduate training after college), race/ethnicity (White non-Hispanic, Black non-Hispanic, Hispanic, 54 Multiracial non-Hispanic, Other non-Hispanic), pre-pregnancy BMI (kg/m 2 ), parity (first born, ≥ second born, missing), maternal hypertensive disorders of pregnancy (hypertensive [defined as having one or more of the following: preeclampsia-eclampsia, chronic hypertension, chronic hypertension with pre-eclampsia superimposed, gestational hypertension] or non-hypertensive), and infant sex (female, male). Minimal sufficient adjustment sets identified in DAGs for both GA at birth and birthweight included maternal age, pre-pregnancy BMI, parity, race/ethnicity, smoking during pregnancy, maternal hypertensive disorders of pregnancy, and socioeconomic status (i.e., income, education). Infant sex was additionally adjusted for in all full GA at birth models since it is an important predictor of GA at birth; however, since BW for GA z- scores were sex-specific, BW for GA z-score models did not include infant sex as a covariate. Pre- pregnancy BMI was calculated using participant reported pre-pregnancy weight and standing height measured by study staff in the first study visit using a commercial stadiometer (Perspectives Enterprises model P-AIM-101). Maternal age, income, education, race/ethnicity, and parity were collected via interviewer administered questionnaires in the participant’s preferred language (English and/or Spanish). Maternal hypertensive disorders of pregnancy were abstracted from maternal prenatal records from health care providers. Sample collection variables, including season at sample collection (winter [December- February], spring [March-May], summer [June-August], autumn [September-November]) and gestational age at sample collection (in weeks), were collected at the time of urine sample collection. Maternal smoking during pregnancy was also considered as a potential covariate; however, given the low frequency of smoking reported among participants (~1.9%), prenatal smoking was not adjusted for in models. However, sensitivity analyses were performed excluding maternal participants who smoked during pregnancy. Although not identified as a potential covariate a priori based on previous literature, other sensitivity analyses included additionally adjusting for gestational diabetes and delivery method to assess the robustness of our results. Gestational diabetes diagnoses and delivery method were abstracted from maternal prenatal records from health care providers. 55 Statistical Analysis Descriptive characteristics were calculated using means and frequencies. OPE metabolite distributions were explored using boxplots, geometric means, medians, percentile distributions, and metabolite detect frequencies. Since OPE metabolite distributions were right skewed, Kruskal Wallis tests were conducted to evaluate univariate associations between categorical covariates (i.e., education, income, race/ethnicity, parity, maternal hypertensive disorders during pregnancy, infant sex, season of sample collection) and OPE concentrations. Spearman correlations were performed to evaluate associations between continuous covariates (maternal age, pre-pregnancy BMI, and gestational age at sample collection) and OPEs as well as correlations between OPE metabolites. Given that OPE concentrations were right skewed, 18 continuous OPE metabolites were natural log transformed prior to modeling. Unadjusted and adjusted linear regressions were performed to assess the relationship between individual prenatal OPEs and infant gestational age at birth and BW for GA z-scores. OPE metabolites with detect frequencies >60% were analyzed continuously (DPHP, DNBP+DIBP, BDCIPP, BCEP, BBOEP) and those with a detect frequency <60% were analyzed dichotomously as detect versus non-detect (BCIPP, BEHP, BMPP, DPRP), consistent with prior literature. 62-64 Model parameter estimates were back transformed and scaled to a doubling in OPE metabolite exposure. Modeling assumptions for all linear regressions, including linearity, homoscedasticity, and normality, were examined (and met) using scatterplots and histograms of model residuals, and all models assessed for influential outliers. A statistical interaction between each OPE metabolite and infant sex was also tested within independent models, with a p-value<0.05 defined as a statistically significant interaction. Regardless of whether statistical interaction terms were significant, all models were stratified by infant sex, given prior evidence of sex-specific effects of OPE metabolites on birth outcomes. Generalized Additive Models (GAMs) with a smoothing term for exposure variables were also performed to evaluate possible non-linear associations between continuously analyzed OPE metabolites and birth outcomes using the R package “mgcv.” All GAM models were also stratified by infant sex. 56 Bayesian Kernel Machine Regression (BKMR) using the Gaussian predictive process (GPP) set at 100 knots was performed to evaluate the associations between the OPE mixture and infant birth outcomes. Only metabolites with a detect frequency >60% were included in BKMR models (n=5 metabolites; DPHP, DNBP+DIBP, BDCIPP, BCEP, BBOEP). BKMR is an advanced semi-parametric method that utilizes Gaussian kernel machine regression to estimate the effects of a high-dimensional matrix of predictors (e.g., interrelated environmental exposures) on a health outcome of interest. 65 Advantages of BKMR include the ability to estimate flexible exposure-response functions which may include non-linear effects and correlated exposures, thus accounting for potential confounding from co-exposures. Additionally, BKMR can evaluate all possible synergistic and antagonistic relationships between mixtures components without prior specification. In brief, the BKMR model for the current study is represented by the following equation: 𝑌 𝑖 = ℎ(𝐷𝑃𝐻𝑃 𝑖 , 𝐷𝑁𝐵𝑃 + 𝐷𝐼𝐵𝑃 𝑖 , 𝐵𝐷𝐶𝐼𝑃𝑃 𝑖 , 𝐵𝐶𝐸𝑃 𝑖 , 𝐵𝐵𝑂𝐸𝑃 𝑖 ) + 𝑋 𝑖 𝛽 + 𝜀 𝑖 where 𝑌 𝑖 represents our health outcome of interest (i.e., infant birth outcome) for participant i, ℎ(.) denotes the exposure-response function; 𝛽 represents the vector of coefficients for model covariates (𝑋 𝑖 ), which are modeled parametrically; and 𝜀 represents residuals assumed to be independent, normally distributed, with a common variance. BKMR models for each outcome were stratified by sex to assess possible sex differences in the association between OPE mixtures and infant birth outcomes. Additionally, all OPE metabolites and infant GA at birth were natural log transformed, mean-centered, and standard deviation scaled prior to BKMR modelling. All continuous covariates were mean centered and scaled to one standard deviation. Individual mixture component associations for GA at birth and BW for GA z-scores were ranked using posterior inclusion probabilities (PIPs) to assess the importance of each mixture component exposure in defining the exposure-outcome association. Exposure-response functions for each OPE holding all other exposures constant at their 50 th percentiles were evaluated to assess the associations of individual OPEs, accounting for the rest of the mixture, with infant birth outcomes. The overall effect of simultaneously 57 increasing all exposures in the mixtures was additionally evaluated to assess the relationship between the overall OPE mixture and infant GA at birth and BW for GA z-scores. Since BKMR is a Bayesian approach, common frequentist approaches, such as evaluation of statistical significance and the use of 95% Confidence Intervals, are not applicable. We used 95% credible intervals to determine the uncertainty in each exposure–outcome association; if the 95% credible interval did not span 0, we considered the metabolite or mixture to be associated with the outcome. Possible pairwise interactions between OPE metabolites were also investigated visually for each birth outcome by assessing the association between each OPE metabolite and birth outcome when varying a second OPE metabolite to its 25 th , 50 th , and 75 th percentile (holding all other OPE metabolites at their 50 th percentile) with non-parallel lines indicating possible pairwise interactions. The “bkmr” package in R was used for the BKMR analysis. 66 The Markov Chain Monte Carlo (MCMC) sampler was used to obtain 100,000 posterior samples of model parameters, with the first half of iterations used as burn-in and chains thinned to every 10 th iteration to reduce potential autocorrelation. Trace plots were visually inspected to assess model convergence, with all models indicating occurrence of convergence. BKMR models were assumed to have non-informative prior distributions in primary models, the default specified in the R package. However, since BKMR can be sensitive to the choice of model priors, sensitivity analyses were conducted to evaluate whether results from the primary model were robust to alternative prior assumptions. The parameter b, which controls the smoothness of the exposure-outcome relationship, was varied to a lower degree of smoothness (b=50) and to a higher degree of smoothness (b=1000). Data were managed and linear regression models were analyzed using SAS v9.4 (SAS Institute, Inc., Cary, NC, USA). GAMs and BKMR models were performed using R (v 4.1.0). The significance level for single chemical metabolite models was set at an alpha of 0.05. 58 RESULTS Participant Characteristics Maternal and infant characteristics of the study population are shown in Table 1. Maternal participants were on average 28.9±6.1 years old at study recruitment, had a pre-pregnancy BMI of 28.6±6.7 kg/m 2 , and were predominantly Hispanic (77.9%). More than half of participants had at most a high school education (57.0%) and 47.5% of participants had an annual household income of <$30,000. Very few (1.9%) maternal participants smoked during pregnancy; however, 21.1% had one or more hypertensive disorders of pregnancy. Infants were primarily second born or more (62.0%), with 9.3% born preterm (<37 weeks) and 3.8% born with a low birth weight (≤2,500 grams). Infants were born at an average gestational age of 39.1±1.5 weeks. This subset of the MADRES cohort was similar to the full cohort on key demographic characteristics including income, ethnicity, maternal age, education, infant sex, and recruitment site. Distributions for all measured OPE metabolites in urine are shown in Table 2. Five metabolites had detection frequencies ≥ 60% (DPHP, DNBP+DIBP, BDCIPP, BCEP, BBOEP) and four metabolites had detection frequencies between 23.8% and 53.9% (BCIPP, BEHP, BMPP, DPRP). OPE concentrations ranged from 0.004 ng/mL to 168.00 ng/mL. BDCIPP had the highest median metabolite concentration (1.29 ng/mL) across all OPE metabolites measured in this sample. Figure 1 illustrates Spearman correlations between all measured OPE metabolites in the full sample and stratified by sex. In general, most OPE metabolites were significantly correlated with one another, although correlations were generally weak, with the highest correlations observed between BDCIPP and DPHP in both the full sample (𝜌 = 0.31, p<0.01) and sex-stratified Spearman correlations (female 𝜌 = 0.35, p<0.01; male 𝜌 = 0.26, p<0.01). Compared to a representative US sample of National Health and Nutrition Examination Survey (NHANES) participants (Table S1), MADRES participants’ median BDCIPP concentrations (1.29 ng/mL) and BCEP concentrations (0.53 ng/mL) were higher than NHANES participants’ median BDCIPP (0.87 ng/mL) and BCEP concentrations (0.39 ng/mL). However, MADRES participants’ median DPHP metabolite 59 concentrations (0.77 ng/mL) and BCIPP concentrations (0.18 ng/mL) were similar to NHANES participants’ median DPHP concentrations (0.79 ng/mL) and BCIPP concentrations (0.19 ng/mL). Univariate associations between OPE metabolite concentrations and covariates are shown in Supplemental Tables 2a, 2b, and 2c. Participants whose 3 rd trimester urinary specimens were collected during the summer had higher median concentrations of DPHP (0.88 ng/mL) and BDCIPP (1.99 ng/mL) when compared to samples collected from participants in the winter (median DPHP= 0.63 ng/mL; median BDCIPP=0.99 ng/mL), spring (median DPHP=0.73 ng/mL; median BDCIPP= 0.92 ng/mL), and autumn (median DPHP=0.86 ng/mL; median BDCIPP=1.36 ng/mL) seasons. Additionally, specimens collected during the summer and autumn had higher median concentrations of BCEP (summer=0.62 ng/mL; autumn= 0.62 ng/mL) when compared to those collected in the winter (median BCEP=0.36 ng/mL) and spring (median BCEP=0.49 ng/mL) seasons. Participants who were underweight prior to the pregnancy had higher median concentrations of BCEP (0.79 ng/mL) when compared to normal weight (0.42 ng/mL), overweight (0.70 ng/mL), and obese (0.40 ng/mL) participants. Participants who reported not knowing their annual household income had lower median DPHP levels (0.67 ng/mL) compared to those reporting an annual household income of <$30,000 (0.85 ng/mL) and ≥$30,000 (0.85 ng/mL). Maternal participants with any hypertensive disorder of pregnancy had significantly higher median BDCIPP metabolite concentrations (1.49 ng/mL) than those with no hypertensive disorders of pregnancy (1.18 ng/mL). Significant differences were observed between income levels and the detection of BEHP, with higher detect frequencies of BEHP among participants with an annual household income <$30,000 (31.50%) when compared to those with an income of ≥$30,000 (22.00%) and those who reported not knowing their income (18.18%). Higher detect frequencies of BMPP were observed in mothers who had male infants (44.66%) compared to female infants (33.02%). Individual Metabolite Associations with Birthweight and Gestational Age at Birth As shown in Tables 3 and 4, no significant associations were observed between maternal OPE metabolites and infant birth outcomes in either unadjusted or covariate-adjusted linear regression models for the full sample. Similarly, there were no significant associations between OPEs and infant birth 60 outcomes when using GAMs, but there was some evidence of non-linear patterns (Figure 2 and Figure 3). In particular, the pattern between DNBP+DIBP and GA at birth and BBOEP and GA at birth and BW for GA had an inverted U-shape. When we examined sex-specific associations using linear regression models, we found a significant interaction between prenatal BDCIPP concentrations and infant sex on GA at birth (p=0.04). In stratified models, a significant inverse association between BDCIPP and GA at birth was observed among males (𝛽 = -0.12; 95% CI: -0.24, -0.01) but not among females (𝛽 = -0.00001; 95% CI: -0.09, 0.09). Although the interaction between DNBP+DIBP and sex was not statistically significant (p=0.10), there was a marginally significant inverse association observed between DNBP+DIBP and GA at birth among female infants (𝛽 = -0.20; 95% CI: -0.40, 0.00). The associations observed in linear regression models between BDCIPP and GA at birth were consistent to those observed in sex-stratified GAMs (Figure 2), with a significant and linear inverse association between higher BDCIPP concentrations and earlier GA at birth among males (p=0.04). However, there was evidence of a non-linear and inverted U-shape association between DNBP+DIBP and GA at birth among female infants (p= 0.01), with a significantly earlier GA at birth at higher concentrations of DNBP+DIBP (p=0.03). There was also some evidence of non-linear patterns between BBOEP and GA at birth among males (p= 0.03) and DNBP+DIBP and BW for GA z-scores among females (p=0.03) and males (p=0.03), although the associations between each metabolite and infant birth outcome in both models were not statistically significant. In sensitivity analyses excluding participants who smoked during pregnancy (n=8) and models additionally adjusting for gestational diabetes (n= 37; 8.8%), exposure effect estimates in linear regression models (Table S3–S6) and associations observed using GAMs (Figure S2–S5) were not meaningfully changed for either outcome (GA at birth and BW for GA z-scores). Similarly, models additionally adjusting for delivery method yielded similar effect estimates to the primary results in both linear regression models (Table S7–S8) and associations when using GAMs (Figure S6–S7). 61 Associations of OPE mixtures with BW for GA z-scores and Gestational Age at Birth Table 5 shows PIP ranks for both the full sample and sex-stratified mixtures models to quantify the importance of each OPE metabolite in the joint mixture effects on each birth outcome. In the overall sample, DNBP+DIBP had the highest PIP for GA at birth while BBOEP had the highest PIP for BW for GA z- scores. In models stratified by infant sex, among female infants, DNBP+DIBP had the highest PIP for the GA at birth and BW for GA z-score models. Among male infants, BDCIPP had the highest PIP for the GA at birth model and DNBP+DIBP had the highest PIP for the BW for GA z-score model. Relationships between each metabolite and GA at birth, fixing other metabolites at their median values and adjusting for key covariates, are shown in Figure 4A. We found an inverse linear association between BDCIPP and GA at birth and an inverse, somewhat linear association between DPHP and BBOEP and GA at birth, with a slight increase in GA at birth at moderate concentrations of this metabolite. For DNBP+DIBP and GA at birth, there was an inverted U-shaped association, consistent with the non-linear pattern observed between DNBP+DIBP and GA at birth when using GAMs. However, effect estimates evaluating the difference in GA at birth for a change in the specified metabolite from the 25 th to the 75 th percentile, holding all other metabolites in the mixture at their median and adjusting for key covariates, had 95% Credible Intervals that spanned 0 (Table S9). The cumulative association between the overall metabolite mixture and GA at birth had a non-monotonic inverted U-shaped association, with lower GA at birth at higher metabolite levels when compared to their median values, and 95% Credible Intervals which did not cross 0 from the 80 th to the 95 th percentile (Figure 4B). The relationship between the OPE mixture and GA at birth varied by infant sex (Figure 5 and 6). In models for female infants, the exposure-response function for DPHP appeared to be inverse and linear whereas the exposure-response function for BDCIPP appeared to be positive and linear (Figure 5A). The association between DNBP+DIBP and GA at birth followed an inverted U-shape, with higher GA at birth values observed for moderate DNBP+DIBP concentrations. However, all effect estimates evaluating the difference in GA at birth for a change in the specified metabolite from the 25 th to the 75 th percentile, holding all other metabolites in the mixture at their median and adjusting for key covariates, had 95% Credible 62 Intervals that spanned 0 (Table S9). At high concentrations, the overall metabolite mixture was associated with a lower GA at birth, with 95% Credible Intervals not crossing 0 when metabolites were set to their 90 th and 95 th percentile when compared to their median values (Figure 5B). Similar to results for female infants, the univariate exposure-response association between BCEP and GA at birth was null for male infants (Figure 6A). However, in contrast to findings for female infants, a marginal inverse association was identified between BDCIPP and GA at birth for males, with a change in log BDCIPP from the 25 th to the 75 th percentile when all other metabolites were set at the median associated with a decrease in GA at birth of -0.11 (-0.26, 0.03) standard deviations (Table S9). A positive association was identified between DNBP+DIBP and GA at birth at low to moderate concentrations and DPHP and BBOEP had an inverse, non-linear association with GA at birth. However, similar to the full sample and female stratified analysis, all metabolite single effect estimates had 95% Credible Intervals that crossed 0 (Table S9). Compared to the overall model, the male stratified cumulative association between the overall metabolite mixture and GA at birth had a similar non-monotonic inverted U-shape, but all 95% Credible Intervals spanned 0 (Figure 6B). For both the full and sex-stratified BW for GA z-score models (Figure 4C, 5C, 6C), the univariate exposure-response association between BCEP and BW for GA at birth was null. In models for female infants, DPHP and BW for GA z-scores had a slight non-linear, inverse association which was similar to the linear inverse association observed among all infants but varied from the inverted U-shape association observed for males. In models for all infants, associations between DNBP+DIBP and BW for GA z-scores and BBOEP and BW for GA z-scores had an inverted U-shape, with a more pronounced shape among females but a U-shape association between DNBP+DIBP and BW for GA z-scores among males and a J shaped association between BBOEP and BW for GA z-scores. Among male models, an inverse, non-linear association was identified between BDCIPP and BW for GA z-scores; however, an inverted U-shape association was observed among females. All individual metabolite associations (holding the rest of the metabolites constant at their median values) had single effect estimates with 95% Credible Intervals that spanned the null when the specified metabolite was changed from the 25 th to the 75 th percentile (Table S10). 63 Similarly, there was no evidence for a cumulative association between the full and sex-stratified metabolite mixtures and BW for GA z-scores (Figure 4D, 5D, 6D). Pairwise interactions between OPE metabolites and birth outcomes were visually explored and several potential interactions were identified (Figure S8). In the full sample, for the GA at birth model, possible interactions were visually identified between BDCIPP and DNBP+DIBP, such that the inverse association between BDCIPP and GA at birth appeared to be stronger at higher levels of DNBP+DIBP (Figure S8A). Additional possible interactions visually identified between metabolites and GA at birth models included BBOEP and DNBP+DIBP, and DPHP and DNBP+DIBP. No pairwise interactions were visually identified for GA at birth in models for female infants (Figure S8C). In models for male infants, a potential pairwise interaction between BDCIPP and DPHP was visually observed, with the inverse association between BDCIPP and GA at birth being stronger at higher levels of DPHP (Figure S8E). The positive association between DNBP+DIBP and GA at birth also appeared to vary by levels of DPHP, such that the association was slightly attenuated at higher levels of DPHP. In the full sample, the inverse association between BDCIPP and BW for GA z-scores was stronger at lower quantiles of BCEP (Figure S8B). The inverse association between BDCIPP and BW for GA z-scores was stronger at higher quantiles of DNBP+DIBP. In models for female infants, the association between BDCIPP and BW for GA z-score varied by levels of DNBP+DIBP (Figure S8D). In models for male infants, the positive association between BBOEP and BW for GA z-score was stronger at higher levels of DPHP and a stronger inverse association between BCEP and BW for GA z-score was observed higher levels of DPHP (Figure S8F). Additionally, the association between BBOEP and BW for GA z-scores among males was stronger at higher quantiles of DNBP+DIBP and DPHP but attenuated at higher quantiles of BDCIPP. Results from sensitivity analyses excluding mothers who reported smoking during pregnancy (N=8), were consistent with the primary analysis, with a slight attenuation in associations (Figure S9-S10). Alternative prior assumptions were also explored at lower (b=50) and higher (b=1000) degrees of smoothness. Results from models assuming a lower degree of smoothness were very similar to primary results and results were inversely linear for models assuming a higher degree of smoothness (Figure S11- 64 S16). These results were consistent with the expected results when relaxing and constricting degrees of smoothness, suggesting our results were robust. Sensitivity analyses additionally adjusting for gestational diabetes (Figure S17-S18) and delivery method (Figure S19-S20) were similarly consistent with the primary analysis, with some stronger associations observed between higher OPE mixtures and earlier GA at birth in males when adjusting for delivery method. DISCUSSION We found evidence that prenatal OPE exposures adversely impact infant GA at birth in sex specific ways when using both traditional single exposure models and a flexible environmental mixture modeling approach, among a sample of predominately low-income, Hispanic participants. Specifically, we found that higher prenatal BDCIPP concentrations were associated with an earlier GA at birth among males in single exposure models. This metabolite also ranked as the highest predictor for GA at birth among males when using a mixture modeling approach. Among female infants, higher exposures to prenatal DNBP+DIBP concentrations were associated with earlier GA at birth when using flexible single exposure models, and DNBP+DIBP similarly ranked as the most important predictor for GA at birth in mixture models, with a possible non-linear association identified. Additionally, higher cumulative OPE metabolite concentrations were associated with an earlier GA at birth, both in the full sample and among females only. There was no evidence of an association between OPEs and BW for GA. Overall, our results stress the importance of considering sex specific impacts of prenatal OPE exposures on children’s health and additionally underline the importance of evaluating the impacts of OPE exposures as a mixture on birth outcomes. The few epidemiological studies that have evaluated sex specific effects on associations between prenatal OPE exposures and infant birth outcomes have shown conflicting results. A case-control study published by Luo et al. in 2020 among women in Wuhan, China found that third trimester DPHP levels were associated with significantly increased risk of low birthweight which, after stratification by sex, only remained significant among females. 53 However, median DPHP concentrations among the Wuhan Maternal and Child Healthcare Hospital prospective birth cohort (Table S2) were substantially lower across both cases (0.06 ng/mL) and controls (0.05 ng/mL) when compared to DPHP concentrations in the MADRES 65 sample (0.77 ng/mL), suggesting possible differences in exposure distributions which may result from a variety of factors, including varying sources of exposure, potentially contributing to the discrepancies in results. A similar study by Hoffman et al. published in 2018 conducted in the Pregnancy Infection and Nutrition Study (PIN), a pregnancy cohort of predominately white (79.7%) and college educated (69.6%) women in North Carolina, found that higher prenatal exposure to BDCIPP was associated with increased odds of preterm delivery among female infants. 52 Another study by Crawford et al. published in 2020 on 56 primarily college educated (45.0%) and non-Hispanic white (64.0%) women found that BDCIPP was associated with increased infant length at birth and birthweight in males but DPHP was negatively associated with abdominal circumference at birth in infants overall and female weight. 67 Median BDCIPP concentrations among participants in the Women and Infants Hospital of Rhode Island cohort (1.18 ng/mL) were slightly lower than concentrations in the MADRES cohort (1.29 ng/mL), but were fairly comparable overall; however, these slight differences, along with other demographic and exposure collection differences, may have contributed to the discrepancies in results. A small study published by Kuiper et al. in 2020 among predominately white (53%) and college educated women (60%) enrolled in the Origins of Child Health and Resilience in Development (ORCHARD) pregnancy cohort (n=76) found no significant associations between OPE metabolites collected throughout pregnancy and infant BW for GA z-scores or GA at delivery, although inverse patterns observed between BDCIPP and DPHP and BW for GA z-scores and GA at delivery were similar to our results. 31 Median BDCIPP concentrations among the MADRES study were two-fold higher than concentrations among ORCHARD participants (0.51 ng/mL) but DPHP concentrations were lower among MADRES participants compared to concentrations among ORCHARD participants (1.12 ng/mL). Another study on 340 mother-infant dyads participating in the Health Outcomes and Measures of the Environment (HOME) cohort located in Cincinnati, Ohio found positive associations between 16-week BCEP and 26-week DPHP with gestational age, but inverse associations between BCEP at 16 weeks and birthweight among female newborns and 26 week DNBP and ponderal index at birth among male newborns. 68 Geometric mean concentrations of DPHP (1.82 ng/mL) and BCEP (0.60 ng/mL) among 66 HOME participants were generally higher than geometric mean concentrations observed among MADRES participants but BDCIPP concentrations (0.80 ng/mL) were relatively lower. Our study found sex-specific adverse impacts of prenatal BDCIPP on male GA at birth and DNBP+DIBP on female GA at birth in individual metabolite models. OPE mixtures analyses similarly indicated the strongest influence of BDCIPP on male GA at birth and DNBP+DIBP on female GA at birth. Our mixtures analysis overall showed that higher exposure to the overall OPE mixture (i.e., higher percentiles of exposures) was associated with adverse impacts on infant GA at birth, with more pronounced associations on female infants. Discrepancies in findings across each of these studies may be driven by multiple factors, including the previously discussed differences in exposure distributions across studies, along with varying geographic characteristics, underpowered samples to detect associations, and varying exposure measurement methods (i.e., number of measurements, GA at collection, and varying years of collection) which may contribute to exposure misclassification. For instance, while OPEs were measured at a single timepoint in pregnancy in our study (~ 31.5 weeks), the PIN study (~27 weeks), and the Wuhan Maternal and Child Healthcare Hospital prospective birth cohort (third trimester), OPEs were measured multiple times throughout pregnancy in the Women and Infants Hospital of Rhode Island cohort (~12, 28, 35 weeks), the HOME study (~16 and 26 weeks), and the ORCHARD study (~ 15.3, 22.3, 30.9 weeks), with more exposure misclassification likely in studies with a single timepoint. Additionally, variability in the geographical location and the years samples were collected may further contribute to differences in exposure distributions, given increasing OPE use in previous decades and state/country specific regulations which may impact OPE usage. 32 Although the biological mechanisms that contribute to OPE impacts on birth outcomes are largely unknown, several hypothesized sex-specific mechanisms have been proposed. 18,69 One of the primary underlying mechanisms suggested by growing literature is the sex-specific impacts of OPE disruptions on the endocrine system, particularly thyroid hormones, which play a critical role in fetal development. 14,70,71 Previous epidemiological literature has found positive associations between prenatal DBUP and thyroid stimulating hormone (TSH) in newborns, specifically female infants, with results suggesting partial 67 mediation by oxidative stress on DNA damage and lipid peroxidation. 14 Associations between BDCIPP and lower levels of newborn triiodothyronine and thyroxine have also been observed, with marginal associations between DPHP and DNBP and lower triiodothyronine and thyroxine levels, but no observed sex-specific interactions. 72 However, animal studies in male zebrafish have observed non-monotonic associations between TDCIPP and testosterone, hypothesized to occur via impacts on estrogenic and anti- androgenic activity on zebrafish. 73,74 Observational studies have also observed associations between concentrations of TDCIPP in dust and BDCIPP in urine with altered hormone levels in adult men. 75,76 Additionally, OPE metabolites may also disrupt the hypothalamic pituitary-adrenal (HPA), hypothalamic- pituitary-thyroid (HPT), and hypothalamic pituitary-gonadal (HPG) axes along with nuclear receptors of the endocrine system, including estrogen receptors, androgen receptors, and glucocorticoid receptors, which are involved in important regulatory mechanisms for fetal development and metabolism. 12,77-79 There are biological indications that these effects may be sex specific given the sex differences in infants’ estrogen receptor expression and the estrogen-like effects of OPEs acting on various hormone related pathways. 80,81 The adverse effects of OPEs on placental development is another hypothesized underlying mechanism for the observed sex-specific association between OPEs and birth outcomes. 33 Prior epidemiological studies have found associations between higher OPEs and increased oxidative stress and reduced immunoreactivity of integrin alpha-1 (ITGA1) and vascular endothelial-cadherin (CDH5) in uterine-invading cytotrophoblast (CTB) cells which may result in placental disruptions and pregnancy complications. 33 Although there is limited literature on the sex-specific effects of OPEs on placentation, some studies suggest sex-specific fetal placentation, with varying responses to environmental insults by fetal sex. 82 For example, a study evaluating brominated flame retardants found that placenta samples of male infants had higher concentrations of PBDEs and thyroid levels despite similar maternal serum concentrations across sex. 83 These sex-specific findings were hypothesized to have resulted from sex differences in the chemical uptake of thyroid hormone transporting membrane proteins in placental tissue, impacting maternal supply of thyroid hormones to support fetal growth. 83 Since placental disruptions and pregnancy complications have been associated with adverse birth outcomes, the possible sex-specific 68 impacts of OPEs on placental functioning may contribute to the sex-specific differences in birth outcomes. 82,84 Given the observed non-monotonic shape between the OPE mixture and infant GA at birth and sex differences in our results, further research evaluating endocrine disruption as a possible underlying mechanism is warranted. Different OPEs are applied to a variety of consumer products as flame retardants and plasticizers, resulting in concurrent exposures to multiple OPEs. 19 Mixtures modeling approaches, such as BKMR, provide us with the opportunity to evaluate the co-occurring impacts of multiple OPEs on infant birth outcomes. For instance, we observed various possible pairwise interactions, suggesting possible synergistic associations between different OPE metabolites and infant birth outcomes which merit further research. Although we observed similar associations between sex-specific single exposure models and mixtures models, there were some minor differences in our results in models evaluating all infants. For example, even though we saw possible inverse patterns, we did not observe significant associations between the individual effects of OPE metabolites on GA at birth among all infants in single metabolite models while we observed lower GA at birth among all infants with a higher cumulative OPE metabolite exposure in mixtures models. This slight difference between the individual chemical and the mixture approach highlights the value of using mixtures methods to get a comprehensive understanding of the impacts of OPEs on infant health, as most individuals are exposed simultaneously to multiple OPE metabolites which may be correlated and may interact with one another. 27 The present study has many important strengths. For one, this study was based on a prospective pregnancy cohort with urine specimens collected before the outcomes of interest. Secondly, measures of urinary OPE metabolites were used to assess exposures and are generally considered reliable indicators of OPE exposures. 19 Additionally, the population evaluated in this study is comprised largely of women originating from Latin America, who are historically underrepresented in U.S. biomedical and population health research, therefore advancing opportunities to address environmental health disparities and inform solutions for environmental justice. We also evaluated various OPE metabolites which are understudied, such as DNBP+DIBP, BCEP, and BBOEP, which advances opportunities for risk assessment and 69 subsequent regulations and interventions. An additional strength was the use of a flexible environmental mixture modeling approach to assess the association between mixtures of OPE metabolites and birth outcomes. This study also has some limitations. For one, only single spot urine samples collected during the third trimester were measured to assess OPE exposures throughout pregnancy which might have led to some exposure misclassification given the relatively short half-life of OPEs. 85-88 However, previous studies indicate moderate to good reproducibility for DPHP and BDCIPP levels throughout pregnancy, although literature on the reproducibility of the remaining metabolites is limited. 28,56 Additionally, despite adjusting for many key covariates identified in the literature, residual confounding could still be present, especially for covariates such as secondhand smoke. We were further underpowered to evaluate clinically relevant cut-offs of our birth outcomes dichotomously, such as preterm birth (≤ 37 weeks) and low birthweight (≤ 2500 grams) as the number of babies born with these conditions was relatively small. Although a big strength of our cohort’s demographic characteristics is the ability to elucidate possible disparities in the impacts of OPEs on early health outcomes, this may limit the generalizability of our results. The timing of OPE collection (mean GA at collection: 31.5 weeks) was an additional limitation of this study given that it may have resulted in survival bias through the exclusion of some extremely preterm (<28 weeks) and very preterm (28-32 weeks) births, since they would not have reached this sample collection timepoint prior to delivery. However, it is important to note that extremely preterm births and very preterm births are particularly rare among the full MADRES cohort (extremely preterm=0.3%; very preterm= 0.9%), minimizing some of our concerns regarding the potential impacts of survival bias on our results. Additionally, if higher exposures to OPEs truly result in adverse impacts on GA at birth, the timing of OPE collection would have likely biased results towards the null since infants most impacted would have been excluded from the analysis. 70 CONCLUSIONS Our results suggest sex-specific adverse effects of exposure to OPEs on infant GA at birth. Since infant GA at birth is an important predictor of lifelong health, this study highlights the potential for long- lasting health implications of toxic environmental exposures during pregnancy and the importance of chemical reduction strategies to support healthy fetal development. Future research aimed at investigating the underlying mechanisms between OPE exposures and GA at birth is warranted. 71 Figure 1: Spearman Correlations of Organophosphate Ester Metabolites (ng/mL) in Third Trimester Maternal Urine A. Full Sample (N=421) B. Females Only (N=215) C. Males Only (N=206) −1 − 0.8 − 0.6 − 0.4 − 0.2 0 0.2 0.4 0.6 0.8 1 DPHP DNBP+DIBP BDCIPP BCEP BBOEP 0.16 0.31 0.11 0.08 0.17 0.14 0.16 0.12 0.15 0.08 −1 − 0.8 − 0.6 − 0.4 − 0.2 0 0.2 0.4 0.6 0.8 1 DPHP DNBP+DIBP BDCIPP BCEP BBOEP 0.14 0.35 0.12 0.14 0.22 0.12 0.25 0.22 0.25 0.07 −1 − 0.8 − 0.6 − 0.4 − 0.2 0 0.2 0.4 0.6 0.8 1 DPHP DNBP+DIBP BDCIPP BCEP BBOEP 0.18 0.26 0.09 0.02 0.13 0.16 0.07 0.02 0.04 0.09 72 Figure 2: Associations Between Prenatal OPE Urinary Metabolite Concentrations (ng/mL) and Gestational Age at Birth, Using Generalized Additive Models (GAMs) Models adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, infant sex, and maternal hypertensive disorders of pregnancy. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. *Significant association between metabolite and GA at birth; † Significant non-linearity. † † 73 Figure 3: Associations Between Prenatal OPE Metabolite Concentrations (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Using Generalized Additive Models (GAMs) Models adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, and maternal hypertensive disorders of pregnancy. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. *Significant association between metabolite and GA at birth; †Significant non-linearity. † † 74 Figure 4: OPE Metabolite Mixtures (ng/mL) and Infant Gestational Age at Birth (weeks) and Birthweight for Gestational Age (BW for GA) Z Scores in Full Models, Using BKMR (N=421) GA at Birth A. B. BW for GA Z Scores C. D. Figure 4 includes the univariate relationship between each metabolite and birth outcome, while other metabolites are fixed at their medians, and a rug plot showing the distribution of the specified metabolite along the x-axis of each panel (column 1) and the cumulative metabolite mixture results showing the estimated difference in the birth outcome when setting all metabolites to the percentile specified on the x-axis compared with setting all metabolites to their median values (column 2). All GA at birth outcome models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, infant sex, and maternal hypertensive disorders of pregnancy. All BW for GA z score outcome models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, and maternal hypertensive disorders of pregnancy. OPE metabolites and GA at birth were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were mean-centered and standard deviation scaled. Note: BKMR, Bayesian Kernel Machine Regression; OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 75 Figure 5: OPE Metabolite Mixtures (ng/mL) and Infant Gestational Age at Birth (weeks) and Birthweight for Gestational Age (BW for GA) Z Scores in Female Stratified Models, Using BKMR (N=215) GA at Birth A. B. BW for GA Z Scores C. D. Figure 5 includes the univariate relationship between each metabolite and birth outcome, while other metabolites are fixed at their medians, and a rug plot showing the distribution of the specified metabolite along the x-axis of each panel (column 1) and the cumulative metabolite mixture results showing the estimated difference in the birth outcome when setting all metabolites to the percentile specified on the x-axis compared with setting all metabolites to their median values (column 2). All GA at birth outcome models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, infant sex, and maternal hypertensive disorders of pregnancy. All BW for GA z score outcome models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, and maternal hypertensive disorders of pregnancy. OPE metabolites and GA at birth were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were mean-centered and standard deviation scaled. Note: BKMR, Bayesian Kernel Machine Regression; OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 76 Figure 6: OPE Metabolite Mixtures (ng/mL) and Infant Gestational Age at Birth (weeks) and Birthweight for Gestational Age (BW for GA) Z Scores in Male Stratified Models, Using BKMR (N=206) GA at Birth A. B. BW for GA Z Scores C. D. Figure 6 includes the univariate relationship between each metabolite and birth outcome, while other metabolites are fixed at the median, and a rug plot showing the distribution of the specified metabolite along the x-axis of each panel (column 1) and the cumulative metabolite mixture results showing the estimated difference in the birth outcome when setting all metabolites to the percentile specified on the x-axis compared with setting all metabolites to their median values (column 2). All GA at birth outcome models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, infant sex, and maternal hypertensive disorders of pregnancy. All BW for GA z score outcome models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, and maternal hypertensive disorders of pregnancy. OPE metabolites were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were mean centered and standard deviation scaled. Note: BKMR, Bayesian Kernel Machine Regression; OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 77 Table 1: Participant Characteristics (N=421) Mean (SD)/Freq(%) Maternal Characteristics Age (years) 28.9 (6.1) Education ≤High School Some College or College Graduate Graduate School 240 (57.0%) 152 (36.1%) 29 (6.9%) Income Don’t Know Less than $30,000 $30,000 or more 121 (28.7%) 200 (47.5%) 100 (23.8%) Race and Ethnicity White, non-Hispanic Black, non-Hispanic Hispanic Multiracial, non-Hispanic Other, non-Hispanic 29 (6.9%) 49 (11.6%) 328 (77.9%) 5 (1.2%) 10 (2.4%) Personal Smoking During Pregnancy No Yes 413 (98.1%) 8 (1.9%) Hypertensive Non-Hypertensive Hypertensive Pre-eclampsia Chronic Hypertension Chronic Hypertension and Pre-eclampsia Gestational Hypertension 332 (78.9%) 89 (21.1%) 39 (9.3%) 10 (2.4%) 12 (2.9%) 28 (6.7%) Pre-Pregnancy BMI (kg/m 2 ) 28.6 (6.7) Infant Characteristics Mean (SD)/Freq(%) Sex Female Male 215 (51.1%) 206 (48.9%) Birth Order First Born Second or more Missing 146 (34.7%) 261 (62.0%) 14 (3.3%) Gestational Age at Birth (weeks) 39.1 (1.5) Birthweight (grams) 3302.9 (475.4) BW for GA z-score -0.1 (1.0) Preterm Birth (<37 weeks) Yes No 39 (9.3%) 382 (90.7%) Low Birthweight (<2500 grams) Yes No 16 (3.8%) 405 (96.2%) 78 Table 2: Distribution of Specific Gravity Adjusted OPE Concentrations (ng/mL) in Maternal Urine (N=421) a Percentiles Distributions Metabolite 25th 50th 75th Min-Max Geometric Mean Detect Frequency LOD (ng/mL) DPHP 0.47 0.77 1.46 0.12-25.59 0.88 99.76% 0.0281 DNBP+DIBP 0.12 0.17 0.26 ND-3.01 0.19 97.62% 0.0441 BDCIPP 0.61 1.29 2.29 ND-68.00 1.08 94.77% 0.0174 BCEP 0.03 0.53 1.62 ND-168.00 0.31 68.41% 0.0200 BBOEP 0.02 0.04 0.07 ND-1.17 0.04 64.85% 0.0199 BCIPP ND 0.18 0.77 ND-40.56 0.13 53.92% 0.0204 BMPP ND 0.01 0.04 ND-0.69 0.02 38.72% 0.0115 BEHP ND ND 0.04 ND-4.42 0.03 25.42% 0.0170 DPRP ND ND 0.05 ND-2.85 0.03 23.75% 0.0278 a Metabolite concentrations below the LOD have been imputed (LOD/√2) and specific gravity adjusted. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate; LOD, Limit of Detection; ND, Not Detected. 79 Table 3: Associations Between Individual OPE Urinary Metabolites (ng/mL) and Gestational Age at Birth (weeks) Full Unadjusted Models 𝛽 (95% CI) (N=421) Full Adjusted Models a 𝛽 (95% CI) (N=421) Metabolite and Infant Sex Interaction P- value (N=421) Female Only Adjusted Models b 𝛽 (95% CI) (N=215) Male Only Adjusted Models b 𝛽 (95% CI) (N=206) DPHP† -0.09 (-0.20, 0.02) -0.09 (-0.20, 0.02) 0.97 -0.09 (-0.23, 0.05) -0.07 (-0.26, 0.12) DNBP+DIBP† -0.04 (-0.19, 0.10) -0.05 (-0.19, 0.10) 0.10 -0.20 (-0.40, 0.00) 0.12 (-0.11, 0.36) BDCIPP† -0.06 (-0.13, 0.01) -0.06 (-0.13, 0.01) 0.04* -0.00 (-0.09, 0.09) -0.12 (-0.24, - 0.01)* BCEP† -0.00 (-0.04, 0.04) 0.00 (-0.04, 0.04) 0.32 -0.02 (-0.08, 0.03) 0.01 (-0.05, 0.08) BBOEP† -0.07 (-0.18, 0.03) -0.07 (-0.17, 0.03) 0.87 -0.08 (-0.22, 0.07) -0.04 (-0.20, 0.12) BCIPP Non-detect Detect REF 0.18 (-0.11, 0.46) REF 0.14 (-0.14, 0.41) 0.32 REF -0.01 (-0.38, 0.36) REF 0.25 (-0.18, 0.68) BMPP Non-detect Detect REF -0.17 (-0.46, 0.12) REF -0.12 (-0.41, 0.16) 0.85 REF -0.11 (-0.50, 0.27) REF -0.09 (-0.53, 0.35) BEHP Non-detect Detect REF -0.04 (-0.36, 0.29) REF 0.04 (-0.28, 0.36) 0.78 REF 0.06 (-0.38, 0.50) REF 0.02 (-0.47, 0.52) DPRP Non-detect Detect REF -0.00 (-0.34, 0.33) REF -0.00 (-0.33, 0.32) 0.40 REF -0.13 (-0.57, 0.31) REF 0.22 (-0.30, 0.73) a Model adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, infant sex, maternal hypertensive disorders during pregnancy. b Model adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, maternal hypertensive disorders of pregnancy. †Beta estimate back transformed to a doubling in exposure. Note: OPE, Organophosphate Ester; CI, Confidence Interval; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di- isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. 80 Table 4: Associations Between Individual OPE Urinary Metabolites (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores Full Unadjusted Models 𝛽 (95% CI) (N=421) Full Adjusted Models a 𝛽 (95% CI) (N=421) Metabolite and Infant Sex Interaction P- value (N=421) Female Only Adjusted Models a 𝛽 (95% CI) (N=215) Male Only Adjusted Models a 𝛽 (95% CI) (N=206) DPHP† 0.01 (-0.07, 0.08) -0.01 (-0.09, 0.06) 0.40 -0.03 (-0.13, 0.08) 0.03 (-0.09, 0.15) DNBP+DIBP† 0.04 (-0.06, 0.15) 0.06 (-0.05, 0.16) 0.62 0.01 (-0.15, 0.17) 0.10 (-0.04, 0.25) BDCIPP† -0.03 (-0.08, 0.02) -0.03 (-0.07, 0.02) 0.48 -0.00 (-0.07, 0.07) -0.05 (-0.12, 0.03) BCEP† 0.00 (-0.03, 0.03) 0.01 (-0.02, 0.04) 0.79 0.01 (-0.03, 0.06) -0.00 (-0.05, 0.04) BBOEP† 0.06 (-0.02, 0.13) 0.04 (-0.03, 0.11) 0.69 0.03 (-0.08, 0.14) 0.07 (-0.03, 0.17) BCIPP Non-detect Detect REF 0.00 (-0.19, 0.20) REF -0.02 (-0.22, 0.17) 0.28 REF 0.11 (-0.17, 0.40) REF -0.15 (-0.42, 0.12) BMPP Non-detect Detect REF 0.10 (-0.10, 0.30) REF 0.06 (-0.13, 0.26) 0.19 REF 0.22 (-0.07, 0.52) REF -0.05 (-0.33, 0.22) BEHP Non-detect Detect REF -0.06 (-0.29, 0.16) REF -0.09 (-0.32, 0.13) 0.30 REF -0.24 (-0.58, 0.10) REF 0.02 (-0.29, 0.33) DPRP Non-detect Detect REF 0.11 (-0.12, 0.34) REF 0.10 (-0.13, 0.32) 0.65 REF 0.04 (-0.30, 0.38) REF 0.16 (-0.17, 0.48) a Model adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, maternal hypertensive disorders of pregnancy. †Beta estimate back transformed to a doubling in exposure Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. 81 Table 5: Posterior Inclusion Probabilities (PIPs) for OPE Urinary Metabolites and Gestational Age at Birth and Birthweight for Gestational Age (BW for GA) Z-scores Across Full and Sex- Stratified BKMR Mixture Models Full Model (N=421) Sex-Stratified Models GA at Birth BW for GA Z Score GA at Birth BW for GA Z Score Female (N=215) Male (N=206) Female (N=215) Male (N=206) DPHP 0.37 0.26 0.23 0.37 0.26 0.37 DNBP+DIBP 0.41 a 0.32 0.82 a 0.39 0.33 a 0.42 a BDCIPP 0.33 0.34 0.14 0.54 a 0.25 0.41 BCEP 0.16 0.28 0.20 0.26 0.26 0.25 BBOEP 0.38 0.40 a 0.20 0.46 0.29 0.41 a Highest value. Note: OPE, Organophosphate Ester; Bayesian Kernel Machine Regression (BKMR); DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 82 SUPPLEMENTAL MATERIALS Supplemental Figure 1a: Directed Acyclic Graph (DAG) for Associations Between Prenatal OPE Urinary Metabolites and Gestational Age (GA) at Birth Directed Acyclic Graph (DAG) used to identify potential confounders and precision variables. The DAG was created using DAGitty. Green ovals represent exposures or predictors of the exposure, pink ovals represent potential confounders, and blue ovals represent the outcome or predictors of outcome. Note: OPE, Organophosphate Ester; SES, Socioeconomic Status; BMI, Body Mass Index; GA, Gestational Age. 83 Supplemental Figure 1b: Directed Acyclic Graph (DAG) for Associations Between Prenatal OPE Urinary Metabolites and Birthweight Directed Acyclic Graph (DAG) used to identify potential confounders and precision variables. The DAG was created using DAGitty. Green ovals represent exposures or predictors of the exposure, pink ovals represent potential confounders, and blue ovals represent the outcome or predictors of outcome. Note: OPE, Organophosphate Ester; SES, Socioeconomic Status; BMI, Body Mass Index; GA, Gestational Age. 84 Supplemental Figure 2: Associations Between Prenatal OPE Urinary Metabolite Concentrations (ng/mL) and Gestational Age (GA) at Birth (weeks), Using Generalized Additive Models (GAMs) (Non-smoking Participants Only (n=413)) Models adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, infant sex, and maternal hypertensive disorders of pregnancy. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di- isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. *Significant association between metabolite and GA at birth; †Significant non-linearity. † † † 85 Supplemental Figure 3: Associations Between Prenatal OPE Metabolite Concentrations (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Using Generalized Additive Models (GAMs) (Non-smoking Participants Only (n=413)) Models adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, and maternal hypertensive disorders of pregnancy. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. *Significant association between metabolite and GA at birth; †Significant non-linearity. † † † 86 Supplemental Figure 4: Associations Between Prenatal OPE Metabolite Concentrations (ng/mL) and GA at Birth, Using Generalized Additive Models (GAMs) (Additional Adjustment for Gestational Diabetes Mellitus (n=421)) Models adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, infant sex, maternal hypertensive disorders of pregnancy, gestational diabetes. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. *Significant association between metabolite and GA at birth; † Significant non-linearity. † † 87 Supplemental Figure 5: Associations Between Prenatal OPE Metabolite Concentrations (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Using Generalized Additive Models (GAMs) (Additional Adjustment for Gestational Diabetes Mellitus (n=421)) Models adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre- pregnancy BMI, income, education, infant birth order, maternal hypertensive disorders of pregnancy, gestational diabetes. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. *Significant association between metabolite and GA at birth; † Significant non-linearity. † 88 Supplemental Figure 6: Associations Between Prenatal OPE Metabolite Concentrations (ng/mL) and GA at Birth, Using Generalized Additive Models (GAMs) (Additional Adjustment for Delivery Method (n=420)) Models adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, infant sex, maternal hypertensive disorders of pregnancy, delivery method. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. *Significant association between metabolite and GA at birth; † Significant non-linearity. † † 89 Supplemental Figure 7: Associations Between Prenatal OPE Metabolite Concentrations (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Using Generalized Additive Models (GAMs) (Additional Adjustment for Delivery Method (n=420)) Models adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre- pregnancy BMI, income, education, infant birth order, maternal hypertensive disorders of pregnancy, delivery method. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. *Significant association between metabolite and GA at birth; † Significant non-linearity. † 90 Supplemental Figure 8: Bivariate Associations Between OPE Urinary Metabolite Mixtures (ng/mL) and Gestational Age (GA) at Birth and Birthweight for Gestational Age (BW for GA) Z-Scores, Using Bayesian Kernel Machine Regression (BKMR) Gestational Age at Birth (n=421) Sex-specific BW for GA Z-scores (N=421) A. Overall Model (n=421) B. Overall Model (n=421) C. Females (n=215) D. Females (n=215) E. Males (n=206) F. Males (n=206) Figure S8 shows BKMR bivariate association results for GA at birth (column 1) and sex-specific BW for GA z Scores (column 2). The bivariate association illustrates the association between each OPE metabolite (labelled in the column) and birth outcome (Y axis), while setting a second metabolite (labelled in the row) to its 25 th , 50 th , and 75 th percentile and all other elements to their median. GA at birth models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, infant sex, and maternal hypertensive disorders of pregnancy. Sex specific BW for GA z-score models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, and maternal hypertensive disorders of pregnancy. OPE metabolites and GA at birth were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were mean-centered and standard deviation scaled. Note: BKMR, Bayesian Kernel Machine Regression; OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 91 Supplemental Figure 9: Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Infant Gestational Age (GA) at Birth (weeks) in Full and Sex-Stratified Models, Using Bayesian Kernel Machine Regression (BKMR) (Non-smoking Participants Only (n=413)) Full Sample (n=413) A. B. Females (n=210) C. D. Males (n=203) E. F. Figure S9 includes the univariate relationship between each metabolite and GA at birth, while other metabolites are fixed at the median, and a rug plot of the distribution of the specified metabolite along the x-axis of each panel (column 1) and the cumulative metabolite mixture results showing the estimated difference in GA at birth when setting all metabolites to the percentile specified on the x-axis compared with setting all metabolites to their median values (column 2). All models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre- pregnancy BMI, income, education, infant birth order, infant sex, and maternal hypertensive disorders of pregnancy. OPE metabolites and GA at birth were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were centered, and SD scaled. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 92 Supplemental Figure 10: Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores in Full and Sex-Stratified Models, Using Bayesian Kernel Machine Regression (BKMR) (Non-smoking Participants Only (n=413)) Full Sample (n=413) A. B. Females (n=210) C. D. Males (n=203) E. F. Figure S10 includes the univariate relationship between each metabolite and BW for GA, while other metabolites are fixed at the median, and a rug plot showing the distribution of the specified metabolite along the x-axis of each panel (column 1) and the cumulative metabolite mixture results showing the estimated difference in BW for GA when setting all metabolites to the percentile specified on the x-axis compared with setting all metabolites to their median values (column 2). All models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, and maternal hypertensive disorders of pregnancy. OPE metabolites were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were mean-centered and SD scaled. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di- isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 93 Supplemental Figure 11: Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Gestational Age (GA) at Birth (weeks) Varying the Smoothing Parameter to b=50 and b=1000, Using Bayesian Kernel Machine Regression (BKMR) (Full Model N=421) b=50 b=1000 A. B. D. E. Figure S11 illustrates BKMR results when varying the smoothing parameter from b=50 (column 1) to b=1000 (column 2). This figure includes the univariate relationship between each metabolite and GA at birth, while other metabolites are fixed at the median (Row 1, Panel A and B) and the cumulative metabolite mixture results showing the estimated difference in GA at birth when setting all metabolites to the percentile specified on the x-axis compared with setting all metabolites to their median values (Row2, Panel D and E). All models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, infant sex, and maternal hypertensive disorders of pregnancy. OPE metabolites and GA at birth were natural log-transformed mean centered, and standard deviation scaled. Continuous covariates were mean-centered and standard deviation scaled. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 94 Supplemental Figure 12: Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Gestational Age (GA) at Birth (weeks) Varying the Smoothing Parameter to b=50 and b=1000, Using Bayesian Kernel Machine Regression (BKMR) (Female Stratified N=215) b=50 b=1000 A. B. D. E. Figure S12 illustrates BKMR results when varying the smoothing parameter from b=50 (column 1) to b=1000 (column 2). This figure includes the univariate relationship between each metabolite and GA at birth, while other metabolites are fixed at the median (Row 1, Panel A and B) and the cumulative metabolite mixture results showing the estimated difference in GA at birth when setting all metabolites to the percentile specified on the x-axis compared with setting all metabolites to their median values (Row2, Panel D and E). All models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, and maternal hypertensive disorders of pregnancy. OPE metabolites and GA at birth were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were mean-centered and standard deviation scaled. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 95 Supplemental Figure 13: Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Gestational Age (GA) at Birth (weeks) Varying the Smoothing Parameter to b=50 and b=1000, Using Bayesian Kernel Machine Regression (BKMR) (Male Stratified N=206) b=50 b=1000 A. B. C. D. Figure S13 illustrates BKMR results when varying the smoothing parameter from b=50 (column 1) to b=1000 (column 2). This figure includes the univariate relationship between each metabolite and GA at birth, while other metabolites are fixed at the median (Row 1, Panel A and B) and the cumulative metabolite mixture results showing the estimated difference in GA at birth when setting all metabolites to the quantile specified on the x-axis compared with setting all metabolites to their median values (Row2, Panel D and E). All models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, and maternal hypertensive disorders of pregnancy. OPE metabolites and GA at birth were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were mean-centered and standard deviation scaled. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 96 Supplemental Figure 14: Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Using Bayesian Kernel Machine Regression (BKMR) Varying the Smoothing Parameter to b=50 and b=1000 (Full Model N=421) b=50 b=1000 A. B. D. E. Figure S14 illustrates BKMR results when varying the smoothing parameter from b=50 (column 1) to b=1000 (column 2). This figure includes the univariate relationship between each metabolite and BW for GA, while other metabolites are fixed at the median (Row 1, Panel A and B) and the cumulative metabolite mixture results showing the estimated difference in BW for GA when setting all metabolites to the quantile specified on the x-axis compared with setting all metabolites to their median values (Row2, Panel D and E). All models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, and maternal hypertensive disorders of pregnancy. OPE metabolites were natural log-transformed, mean centered and standard deviation scaled. Continuous covariates were mean-centered and standard deviation scaled. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2- propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 97 Supplemental Figure 15: Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Using Bayesian Kernel Machine Regression (BKMR) Varying the Smoothing Parameter to b=50 and b=1000 (Female Stratified N=215) b=50 b=1000 A. B. E. F. Figure S15 illustrates BKMR results when varying the smoothing parameter from b=50 (column 1) to b=1000 (column 2). This figure includes the univariate relationship between each metabolite and BW for GA, while other metabolites are fixed at the median (Row 1, Panel A and B) and the cumulative metabolite mixture results showing the estimated difference in BW for GA when setting all metabolites to the quantile specified on the x-axis compared with setting all metabolites to their median values (Row2, Panel D and E). All models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, and maternal hypertensive disorders of pregnancy. OPE metabolites were natural log-transformed, mean centered and standard deviation scaled. Continuous covariates were mean-centered and standard deviation scaled. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2- propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 98 Supplemental Figure 16: Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Using Bayesian Kernel Machine Regression (BKMR), Varying the Smoothing Parameter to b=50 and b=1000 (Male Stratified N=206) b=50 b=1000 A. B. E. F. Figure S16 illustrates BKMR results when varying the smoothing parameter from b=50 (column 1) to b=1000 (column 2). This figure includes the univariate relationship between each metabolite and BW for GA, while other metabolites are fixed at the median (Row 1, Panel A and B) and the cumulative metabolite mixture results showing the estimated difference in BW for GA when setting all metabolites to the quantile specified on the x-axis compared with setting all metabolites to their median values (Row2, Panel D and E). All models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, and maternal hypertensive disorders of pregnancy. OPE metabolites were natural log-transformed, mean centered and standard deviation scaled. Continuous covariates were mean-centered and standard deviation scaled. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2- propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 99 Supplemental Figure 17: Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Infant Gestational Age (GA) at Birth (weeks), Using Bayesian Kernel Machine Regression (BKMR), in Full and Sex- Stratified Models (Additionally Adjusting for Gestational Diabetes Mellitus (n=421)) Full Sample (n=421) A. B. Females (n=215) C. D. Males (n=206) E. F. Figure S17 includes the univariate relationship between each metabolite and GA at birth, while other metabolites are fixed at the median, and a rug plot of the distribution of the specified metabolite along the x-axis of each panel (column 1) and the cumulative metabolite mixture results showing the estimated difference in GA at birth when setting all metabolites to the percentile specified on the x-axis compared with setting all metabolites to their median values (column 2). All models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre- pregnancy BMI, income, education, infant birth order, infant sex, maternal hypertensive disorders of pregnancy, and gestational diabetes. OPE metabolites and GA at birth were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were centered, and SD scaled. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 100 Supplemental Figure 18: Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Using Bayesian Kernel Machine Regression (BKMR) in Full and Sex-Stratified Models (Additionally Adjusting for Gestational Diabetes Mellitus (n=421)) Full Sample (n=421) A. B. Females (n=215) C. D. Males (n=206) E. F. Figure S18 includes the univariate relationship between each metabolite and BW for GA z-scores, while other metabolites are fixed at the median, and a rug plot of the distribution of the specified metabolite along the x-axis of each panel (column 1) and the cumulative metabolite mixture results showing the estimated difference in BW for GA z-scores when setting all metabolites to the percentile specified on the x-axis compared with setting all metabolites to their median values (column 2). All models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, maternal hypertensive disorders of pregnancy, gestational diabetes. OPE metabolites and GA at birth were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were centered, and SD scaled. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 101 Supplemental Figure 19: Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Infant Gestational Age (weeks), Using Bayesian Kernel Machine Regression (BKMR) in Full and Sex-Stratified Models (Additionally Adjusting for Method of Delivery (n=421)) Full Sample (n=421) A. B. Females (n=215) C. D. Males (n=206) E. F. Figure S19 includes the univariate relationship between each metabolite and GA at birth, while other metabolites are fixed at the median, and a rug plot of the distribution of the specified metabolite along the x-axis of each panel (column 1) and the cumulative metabolite mixture results showing the estimated difference in GA at birth when setting all metabolites to the percentile specified on the x-axis compared with setting all metabolites to their median values (column 2). All models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre- pregnancy BMI, income, education, infant birth order, infant sex, maternal hypertensive disorders of pregnancy, delivery method. OPE metabolites and GA at birth were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were centered, and SD scaled. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 102 Supplemental Figure 20: Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Using Bayesian Kernel Machine Regression (BKMR) in Full and Sex-Stratified Models (Additionally Adjusting for Method of Delivery (n=421)) Full Sample (n=421) A. B. Females (n=215) C. D. Males (n=206) E. F. Figure S20 includes the univariate relationship between each metabolite and BW for GA z-scores, while other metabolites are fixed at the median, and a rug plot of the distribution of the specified metabolite along the x-axis of each panel (column 1) and the cumulative metabolite mixture results showing the estimated difference in BW for GA z-scores when setting all metabolites to the percentile specified on the x-axis compared with setting all metabolites to their median values (column 2). All models were adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, maternal hypertensive disorders of pregnancy, delivery method. OPE metabolites and GA at birth were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were centered, and SD scaled. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 103 Supplemental Table 1: Median Concentrations (ng/mL) of Urinary OPE Metabolites Across Published Studies Country (City/State) Cohort (N) BDCIPP DPHP BCIPP DNBP + DIBP BCEP BBOEP BMPP BEHP DPRP Adjustment Population (GA at Collection/Age) Reference US (Los Angeles /CA) MADRES (N=421) 1.29 0.77 0.18 0.17 0.53 0.04 0.01 0.01 0.02 SG adjusted Birth cohort (31.5 (SD: 2.0) weeks) Our Study US National Health and Nutrition Examination Survey (NHANES) (n=2666) 0.87 0.79 0.19 - 0.39 - - - - Creatine adjusted Representative US Cohort (>6 years old) Ospina et al., 2018 US (North Carolina) Pregnancy, Infection, Nutrition (PIN) (N=349) 1.85 1.31 ND _ _ _ _ _ _ SG adjusted Birth cohort (27 weeks (range 24-30)) Hoffman et al. 2018 US (Baltimore, MD) ORigins of Child Health And Resilience in Development (ORCHARD) (N=90) 0.51 1.12 ND _ _ _ _ _ _ SG adjustment Birth cohort (30.9 (SD: 2.5) weeks) Kuiper et al., 2020 US (Salinas/CA) Center for the Health Assessment of Mothers and Children of Salinas (CHAMACOS) (N=310) 0.41 0.93 _ _ _ _ _ _ _ SG adjustment Birth cohort (26.0 (2.4) weeks) Castorina et al., 2017 US (Rhode Island) Women and Infants Hospital of Rhode Island (N=56) 1.18 0.93 _ _ 0.31 _ _ _ _ SG adjustment Birth cohort (1 st -3 rd trimester) Crawford et al., 2020 China (Wuhan) Wuhan Maternal and Child Healthcare Hospital (N=213) 0.10 0.24 _ _ _ 0.14 _ _ _ SG adjusted Birth cohort (1 st -3 rd trimester) Luo et al., 2021 US (Boston/Massachusetts) LIFECODES (N= 90) 0.67 0.74 _ _ _ _ _ _ _ SG adjusted Birth cohort (1 st -3 rd trimester) Bommarito, et al., 2021 US (Cincinnati, OH) HOME (n=340) 0.80 1.82 _ 0.26 0.60 _ _ _ _ SG adjusted Birth cohort (16 weeks and 26 weeks) Yang et al., 2023 104 Note: GA, Gestational Age; OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. 105 Supplemental Table 2a: Prenatal OPE Urinary Metabolite Concentrations By Selected Sample Collection Variables (N=421) N DPHP Median a DBNP+DIBP Median a BDCIPP Median a BCEP Median a BBOEP Median a BCIPP N (DF%) b BMPP N (DF%) b BEHP N (DF%) b DPRP N (DF%) b Season of specimen collection Winter (December-February) Spring (March-May) Summer (June-August) Autumn (September-November) p-value 93 91 106 131 0.63 0.73 0.88 0.86 0.02 0.16 0.17 0.18 0.19 0.11 0.99 0.92 1.99 1.36 <0.01 0.36 0.49 0.62 0.62 0.02 0.05 0.04 0.04 0.04 0.41 53 (56.99%) 42 (46.15%) 59 (55.66%) 73 (55.73%) 0.41 38 (40.86%) 34 (37.36%) 41 (38.68%) 50 (38.17%) 0.97 17 (18.28%) 29 (31.87%) 31 (29.25%) 30 (22.90%) 0.12 22 (23.66%) 14 (15.38%) 29 (27.36%) 35 (26.72%) 0.18 GA at Sample Collection 24-30 weeks 30-34 weeks 34-38 weeks p-value 79 286 56 0.75 0.79 0.72 0.93 0.17 0.17 0.20 0.14 1.17 1.29 1.47 0.61 0.26 0.56 0.64 0.31 0.04 0.04 0.05 0.11 42 (53.16%) 156 (54.55%) 29 (51.79%) 0.92 27 (34.18%) 119 (41.61%) 17 (30.36%) 0.19 22 (27.85%) 72 (25.17%) 13 (23.21%) 0.82 16 (20.25%) 73 (25.52%) 11 (19.64%) 0.46 a Kruskal Wallis tests performed. b Pearson Chi-Square test performed. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. 106 Supplemental Table 2b: Prenatal OPE Urinary Metabolite Concentrations By Selected Maternal Characteristics (N=421) N DPHP Median a DBNP+DIBP Median a BDCIPP Median a BCEP Median a BBOEP Median a BCIPP N (DF%) b BMPP N (DF%) b BEHP N (DF%) b DPRP N (DF%) b Maternal Age Q1 (18.28-23.86) Q2 (23.87-28.69) Q3 (28.81-33.26) Q4 (33.32-45.46) p-value 105 105 106 105 0.74 0.84 0.69 0.85 0.65 0.17 0.18 0.17 0.18 0.19 1.46 1.24 1.28 1.14 0.26 0.62 0.38 0.55 0.49 0.23 0.04 0.05 0.04 0.04 0.38 58 (44.76%) 56 (53.33%) 58 (54.72%) 55 (52.38%) 0.98 37 (35.24%) 39 (37.14%) 46 (43.40%) 41 (39.05%) 0.65 26 (24.76%) 26 (24.76%) 25 (23.58%) 30 (28.57%) 0.85 27 (25.71%) 25 (23.81%) 24 (22.64%) 24 (22.86%) 0.95 Race/ethnicity White, non-Hispanic Black, non-Hispanic Hispanic Multiracial, non-Hispanic Other, non-Hispanic p-value 29 49 328 5 10 0.84 0.93 0.75 1.12 0.49 0.74 0.20 0.22 0.17 0.17 0.22 0.56 1.32 1.83 1.25 1.23 1.32 0.26 0.31 0.51 0.56 1.41 0.82 0.06 0.05 0.03 0.04 0.08 0.03 0.08 19 (65.52%) 28 (57.14%) 173 (52.74%) 4 (80.00%) 3 (30.00%) 0.25 c 13 (44.83%) 19 (38.78%) 123 (37.50%) 3 (60.00%) 5 (50.00%) 0.67 c 8 (27.59%) 16 (32.65%) 79 (24.09%) 1 (20.00%) 3 (30.00%) 0.70 c 6 (20.69%) 10 (20.41%) 81 (24.70%) 0 (0.00%) 3 (30.00%) 0.77 c Pre-pregnancy BMI (kg/m 2 ) Underweight Normal Overweight Obese p-value 12 125 132 152 0.69 0.74 0.72 0.87 0.30 0.16 0.17 0.18 0.18 0.76 1.85 1.14 1.29 1.36 0.43 0.79 0.42 0.70 0.40 0.01 0.03 0.04 0.04 0.04 0.72 9 (75.00%) 70 (56.00%) 61 (46.21%) 87 (57.24%) 0.10 6 (50.00%) 47 (37.60%) 50 (37.88%) 60 (39.47%) 0.85 2 (16.67%) 34 (27.20%) 30 (22.73%) 41 (26.97%) 0.71 3 (25.00%) 24 (19.20%) 31 (23.48%) 42 (27.63%) 0.44 Income Don’t Know <$30,000 ≥$30,000 p-value 121 200 100 0.67 0.85 0.85 0.04 0.17 0.17 0.19 0.27 1.29 1.27 1.32 0.65 0.46 0.60 0.46 0.70 0.05 0.04 0.04 0.56 69 (57.02%) 103 (51.50%) 55 (55.00%) 0.61 44 (36.36%) 75 (37.50%) 44 (44.00%) 0.45 22 (18.18%) 63 (31.50%) 22 (22.00%) 0.02 30 (24.79%) 39 (19.50%) 31 (31.00%) 0.08 Education ≤HS Some college-4 years Grad School p-value 240 152 29 0.78 0.77 0.60 0.73 0.17 0.18 0.20 0.54 1.28 1.32 1.11 0.90 0.55 0.49 0.80 0.80 0.04 0.04 0.04 0.83 125 (52.08%) 81 (53.29%) 21 (72.41%) 0.11 96 (40.00%) 54 (35.53%) 13 (44.83%) 0.53 57 (23.75%) 43 (28.29%) 7 (24.14%) 0.60 61 (25.42%) 32 (21.05%) 7 (24.14%) 0.61 Hypertensive Disorders of Pregnancy No Yes p-value 332 89 0.77 0.74 0.71 0.17 0.18 0.42 1.18 1.49 0.02 0.56 0.42 0.78 0.04 0.04 0.41 181 (54.52%) 46 (51.69%) 0.63 125 (37.65%) 38 (42.70%) 0.39 86 (25.90%) 21 (23.60%) 0.66 77 (23.19%) 23 (25.84%) 0.60 Gestational Diabetes Status Normal 278 0.78 0.17 1.35 0.56 0.04 154 (55.40%) 104 (37.41%) 74 (26.62%) 60 (21.58%) 107 Glucose Intolerant Gestational Diabetes Chronic Diabetes 82 37 24 0.78 0.72 0.69 0.58 0.19 0.18 0.17 0.41 1.25 1.20 1.15 0.31 0.46 0.47 0.59 0.83 0.05 0.04 0.04 0.58 44 (53.66%) 22 (59.46%) 7 (29.17%) 0.09 31 (37.80%) 19 (51.35%) 9 (37.50%) 0.43 22 (26.83%) 8 (21.62%) 3 (12.50%) 0.44 22 (26.83%) 8 (21.62%) 10 (41.67%) 0.14 a Kruskal Wallis tests performed. b Pearson Chi-Square test performed unless 30% of cells expected counts had less than 5 in which case a Fisher Exact Test c was performed. Note: BKMR, Bayesian Kernel Machine Regression; OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3- dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. 108 Supplemental Table 2c: Prenatal OPE Urinary Metabolites By Selected Infant Characteristics (N=421) N DPHP Median a DBNP+DIBP Median a BDCIPP Median a BCEP Median a BBOEP Median a BCIPP N (DF%) b BMPP N (DF%) b BEHP N (DF%) b DPRP N (DF%) b Infant Birth Order Missing First Born Second Born or more p-value 14 146 261 0.91 0.77 0.75 0.39 0.19 0.19 0.17 0.17 0.56 1.38 1.28 0.16 0.62 0.58 0.49 0.33 0.04 0.04 0.04 0.91 8 (57.14%) 82 (56.16%) 137 (52.49%) 0.75 8 (57.14%) 51 (34.93%) 104 (39.85%) 0.22 5 (35.17%) 35 (23.97%) 67 (25.67%) 0.62 3 (21.43%) 36 (24.66%) 61 (23.37%) 0.94 Infant Sex Female Male p-value 215 206 0.75 0.77 0.77 0.18 0.17 0.42 1.23 1.36 0.77 0.58 0.49 0.81 0.04 0.04 0.85 115 (53.49%) 112 (54.37%) 0.86 71 (33.02%) 92 (44.66%) 0.01 55 (25.58%) 52 (25.24%) 0.94 51 (23.72%) 49 (23.79%) 0.99 a Kruskal Wallis tests performed. b Pearson Chi-Square test performed. Note: BKMR, Bayesian Kernel Machine Regression; OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3- dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. 109 Supplemental Table 3: Associations between Prenatal OPE Urinary Metabolite Concentrations (ng/mL) and Gestational Age at Birth in Participants Who Reported No In-Utero Smoking (N=413) Full Adjusted Models a 𝛽 (95% CI) (N=413) Female Only Adjusted Models b 𝛽 (95% CI) (N=210) Male Only Adjusted Models b 𝛽 (95% CI) (N=203) DPHP† -0.09 (-0.20, 0.02) -0.08 (-0.23, 0.06) -0.06 (-0.25, 0.13) DNBP+DIBP† -0.05 (-0.20, 0.10) -0.20 (-0.41, 0.00) 0.12 (-0.12, 0.35) BDCIPP† -0.05 (-0.12, 0.02) 0.01 (-0.08, 0.10) -0.12 (-0.24, -0.00)* BCEP† 0.00 (-0.04, 0.05) -0.02 (-0.08, 0.03) 0.02 (-0.04, 0.09) BBOEP† -0.06 (-0.17, 0.04) -0.08 (-0.23, 0.07) -0.02 (-0.18, 0.15) BCIPP Non-detect Detect REF 0.14 (-0.15, 0.42) REF 0.00 (-0.38, 0.39) 0.24 (-0.19, 0.67) BMPP Non-detect Detect REF -0.14 (-0.43, 0.15) REF -0.11 (-0.51, 0.28) REF -0.13 (-0.57, 0.31) BEHP Non-detect Detect REF 0.07 (-0.26, 0.0.39) REF 0.05 (-0.40, 0.49) REF 0.09 (-0.41, 0.60) DPRP Non-detect Detect REF -0.02 (-0.35, 0.31) REF -0.15 (-0.60, 0.30) REF 0.19 (-0.32, 0.70) a Model adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, infant sex, maternal hypertensive disorders of pregnancy. b Model adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, maternal hypertensive disorders of pregnancy. †Beta estimate back transformed to a doubling in exposure. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2- methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. 110 Supplemental Table 4: Associations between Prenatal Urinary OPE Concentrations (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores Among Participants Who Reported No In-Utero Smoking (N=413) Full Adjusted Models a 𝛽 (95% CI) (N=413) Female Only Adjusted Models a 𝛽 (95% CI) (N=210) Male Only Adjusted Models a 𝛽 (95% CI) (N=203) DPHP† -0.02 (-0.09, 0.06) -0.03 (-0.14, 0.08) 0.03 (-0.09, 0.15) DNBP+DIBP† 0.06 (-0.04, 0.16) 0.01 (-0.15, 0.17) 0.11 (-0.03, 0.26) BDCIPP† -0.03 (-0.08, 0.02) -0.01 (-0.08, 0.06) -0.05 (-0.12, 0.03) BCEP† 0.01 (-0.02, 0.04) 0.01 (-0.03, 0.06) -0.00 (-0.05, 0.04) BBOEP† 0.04 (-0.03, 0.11) 0.03 (-0.09, 0.14) 0.07 (-0.04, 0.17) BCIPP Non-detect Detect REF -0.02 (-0.21, 0.18) REF 0.10 (-0.20, 0.39) REF -0.11 (-0.38, 0.17) BMPP Non-detect Detect REF 0.06 (-0.14, 0.26) REF 0.23 (-0.08, 0.53) REF -0.05 (-0.32, 0.23) BEHP Non-detect Detect REF -0.11 (-0.33, 0.12) REF -0.24 (-0.58, 0.11) REF 0.01 (-0.31, 0.33) DPRP Non-detect Detect REF 0.10 (-0.13, 0.33) REF 0.05 (-0.30, 0.40) REF 0.15 (-0.17, 0.47) a Model adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, maternal hypertensive disorders of pregnancy. †Beta estimate back transformed to a doubling in exposure. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. 111 Supplemental Table 5: Associations Between Prenatal Urinary OPE Concentrations (ng/mL) and Gestational Age (GA) at Birth, Additionally Adjusting for Gestational Diabetes (N=421) Full Adjusted Models a 𝛽 (95% CI) (N=421) Female Only Adjusted Models b 𝛽 (95% CI) (N=215) Male Only Adjusted Models b 𝛽 (95% CI) (N=206) DPHP† -0.10 (-0.21, 0.01) -0.11 (-0.24, 0.02) -0.08 (-0.27, 0.12) DNBP+DIBP† -0.04 (-0.19, 0.10) -0.19 (-0.39, 0.00) 0.11 (-0.12, 0.35) BDCIPP† -0.07 (-0.14, 0.00) -0.01 (-0.10, 0.07) -0.14 (-0.25, -0.02)* BCEP† 0.00 (-0.04, 0.04) -0.02 (-0.07, 0.04) 0.01 (-0.05, 0.08) BBOEP† -0.06 (-0.16, 0.04) -0.06 (-0.20, 0.08) -0.05 (-0.21, 0.12) BCIPP Non-detect Detect REF 0.09 (-0.19, 0.36) REF -0.14 (-0.51, 0.22) REF 0.22 (-0.22, 0.66) BMPP Non-detect Detect REF -0.11 (-0.39, 0.18) REF -0.06 (-0.43, 0.32) REF -0.08 (-0.52, 0.37) BEHP Non-detect Detect REF -0.0007 (-0.32, 0.32) REF 0.01 (-0.41, 0.43) REF -0.01 (-0.52, 0.49) DPRP Non-detect Detect REF 0.05 (-0.27, 0.38) REF -0.05 (-0.48, 0.38) REF 0.25 (-0.27, 0.76) a Model adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, infant sex, maternal hypertensive disorders of pregnancy, gestational diabetes. b Model adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, maternal hypertensive disorders of pregnancy, gestational diabetes. †Beta estimate back transformed to a doubling in exposure. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2- methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. 112 Supplemental Table 6: Associations between Prenatal Urinary OPE Concentrations (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Additionally Adjusting for Gestational Diabetes (N=421) Full Adjusted Models a 𝛽 (95% CI) (N=421) Female Only Adjusted Models a 𝛽 (95% CI) (N=215) Male Only Adjusted Models a 𝛽 (95% CI) (N=206) DPHP† -0.00 (-0.08, 0.07) -0.01 (-0.12, 0.09) 0.03 (-0.09, 0.15) DNBP+DIBP† 0.06 (-0.04, 0.16) 0.02 (-0.14, 0.17) 0.11 (-0.04, 0.26) BDCIPP† -0.02 (-0.07, 0.03) 0.00 (-0.06, 0.07) -0.03 (-0.11, 0.04) BCEP† 0.01 (-0.02, 0.04) 0.01 (-0.03, 0.05) -0.00 (-0.04, 0.04) BBOEP† 0.04 (0.03, 0.11) 0.02 (-0.09, 0.12) 0.06 (-0.04, 0.17) BCIPP Non-detect Detect REF 0.001 (-0.19, 0.19) REF 0.18 (-0.10, 0.47) REF -0.15 (-0.43, 0.12) BMPP Non-detect Detect REF 0.03 (-0.16, 0.23) REF 0.17 (-0.13, 0.47) REF -0.07 (-0.34, 0.21) BEHP Non-detect Detect REF -0.06 (-0.28, 0.16) REF -0.21 (-0.54, 0.13) REF 0.07 (-0.25, 0.38) DPRP Non-detect Detect REF 0.07 (-0.16, 0.29) REF 0.04 (-0.30, 0.38) REF 0.11 (-0.21, 0.43) a Model adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, maternal hypertensive disorders of pregnancy, gestational diabetes. †Beta estimate back transformed to a doubling in exposure. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. 113 Supplemental Table 7: Associations Between Prenatal Urinary OPE Concentrations (ng/mL) and Gestational Age (GA) at Birth, Additionally Adjusting for Delivery Method (N=420) Full Adjusted Models a 𝛽 (95% CI) (N=420) Female Only Adjusted Models b 𝛽 (95% CI) (N=215) Male Only Adjusted Models b 𝛽 (95% CI) (N=205) DPHP† -0.11 (-0.21, 0.00) -0.09 (-0.23, 0.05) -0.12 (-0.31, 0.07) DNBP+DIBP† -0.04 (-0.18, 0.11) -0.20 (-0.41, -0.00) 0.17 (-0.07, 0.40) BDCIPP† -0.06 (-0.13, 0.00) -0.00 (-0.09, 0.08) -0.15 (-0.27, -0.04)* BCEP† -0.00 (-0.04, 0.04) -0.02 (-0.08, 0.03) 0.01 (-0.06, 0.07) BBOEP† -0.07 (-0.17, 0.03) -0.07 (-0.22, 0.07) -0.04 (-0.20, 0.12) BCIPP Non-detect Detect REF 0.12 (-0.16, 0.40) REF -0.05 (-0.42, 0.33) REF 0.23 (-0.20, 0.65) BMPP Non-detect Detect REF -0.09 (-0.37, 0.20) REF -0.07 (-0.46, 0.32) REF -0.06 (-0.49, 0.38) BEHP Non-detect Detect REF 0.04 (-0.28, 0.36) REF 0.03 (-0.41, 0.46) REF 0.03 (-0.47, 0.53) DPRP Non-detect Detect REF -0.01 (-0.34, 0.32) REF -0.09 (-0.54, 0.35) REF 0.15 (-0.36, 0.65) a Model adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, infant sex, maternal hypertensive disorders of pregnancy, delivery method. b Model adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, maternal hypertensive disorders of pregnancy, delivery method. †Beta estimate back transformed to a doubling in exposure. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2- methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. 114 Supplemental Table 8: Associations between Prenatal Urinary OPE Concentrations (ng/mL) and Birthweight for Gestational Age (BW for GA) Z-scores, Additionally Adjusting for Delivery Method (N=420) Full Adjusted Models a 𝛽 (95% CI) (N=420) Female Only Adjusted Models a 𝛽 (95% CI) (N=215) Male Only Adjusted Models a 𝛽 (95% CI) (N=205) DPHP† -0.01 (-0.08, 0.07) -0.02 (-0.13, 0.08) 0.04 (-0.08, 0.16) DNBP+DIBP† 0.06 (-0.04, 0.16) 0.04 (-0.12, 0.20) 0.07 (-0.07, 0.22) BDCIPP† -0.02 (-0.07, 0.02) -0.00 (-0.07, 0.07) -0.04 (-0.11, 0.03) BCEP† 0.01 (-0.02, 0.04) 0.01 (-0.03, 0.05) -0.01 (-0.05, 0.03) BBOEP† 0.03 (-0.04, 0.11) 0.02 (-0.08, 0.03) 0.05 (-0.05, 0.15) BCIPP Non-detect Detect REF -0.02 (-0.22, 0.17) REF 0.11 (-0.18, 0.40) REF -0.15 (-0.41, 0.12) BMPP Non-detect Detect REF 0.05 (-0.14, 0.25) REF 0.23 (-0.07, 0.53) REF -0.02 (-0.30, 0.25) BEHP Non-detect Detect REF -0.09 (-0.31, 0.13) REF -0.22 (-0.55, 0.12) REF -0.02 (-0.33, 0.29) DPRP Non-detect Detect REF 0.09 (-0.14, 0.31) REF 0.02 (-0.32, 0.37) REF 0.17 (-0.15, 0.48) a Model adjusted for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, maternal hypertensive disorders of pregnancy, delivery method. †Beta estimate back transformed to a doubling in exposure. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. 115 Supplemental Table 9: Single Exposure Effect Estimates for Each Prenatal OPE Metabolite in the Bayesian Kernel Machine Regression (BKMR) Mixture and Gestational Age (GA at Birth) Model Metabolite Effect Estimates 95% Credible Interval Full Sample (n=421) DPHP -0.05 -0.19, 0.08 DNBP+DIBP 0.02 -0.10, 0.14 BDCIPP -0.05 -0.15, 0.05 BCEP 0.03 -0.14, 0.19 BBOEP -0.04 -0.18, 0.09 Females (n=215) DPHP -0.07 -0.22, 0.07 DNBP+DIBP -0.02 -0.18, 0.14 BDCIPP 0.03 -0.08, 0.15 BCEP -0.08 -0.31, 0.14 BBOEP -0.05 -0.21, 0.12 Males (n=206) DPHP -0.03 -0.24, 0.18 DNBP+DIBP 0.10 -0.08, 0.28 BDCIPP -0.11 -0.26, 0.03 BCEP 0.04 -0.21, 0.30 BBOEP 0.02 -0.21, 0.25 Effect estimates reflect the difference in GA at birth for a change in the specified metabolite from the 25th to 75th percentile, holding all other metabolites in the mixture at their median values and adjusting for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, infant sex, and maternal hypertensive disorders of pregnancy. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2- propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 116 Supplemental Table 10: Single Exposure Effect Estimates for Each Prenatal OPE Metabolite in the Bayesian Kernel Machine Regression (BKMR) Mixture and Birthweight for Gestational Age (BW for GA) Z-score Model Metabolite Effect Estimates 95% Credible Interval Full Sample (n=421) DPHP -0.02 -0.15, 0.11 DNBP+DIBP 0.04 -0.08, 0.16 BDCIPP -0.04 -0.15, 0.07 BCEP 0.04 -0.15, 0.23 BBOEP 0.13 -0.03, 0.30 Females (n=215) DPHP -0.06 -0.25, 0.14 DNBP+DIBP 0.06 -0.14, 0.26 BDCIPP -0.03 -0.26, 0.19 BCEP 0.07 -0.22, 0.36 BBOEP 0.20 -0.10, 0.50 Males (n=206) DPHP 0.06 -0.15, 0.27 DNBP+DIBP 0.04 -0.14, 0.22 BDCIPP -0.04 -0.18, 0.10 BCEP -0.06 -0.30, 0.19 BBOEP 0.12 -0.12, 0.36 Effect estimates reflect the difference in BW for GA z-score for a change in the specified metabolite from the 25th to 75th percentile, holding all other metabolites in the mixture at their median values and adjusting for recruitment site, maternal age, season of sample collection, gestational age at sample collection, race/ethnicity, pre-pregnancy BMI, income, education, infant birth order, and maternal hypertensive disorders of pregnancy. Note: OPE, Organophosphate Ester; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 117 References 1. Upadhyay RP, Naik G, Choudhary TS, et al. Cognitive and motor outcomes in children born low birth weight: a systematic review and meta-analysis of studies from South Asia. BMC Pediatrics 2019; 19: 35. DOI: 10.1186/s12887-019-1408-8. 2. Whitaker AH, Feldman JF, Lorenz JM, et al. Motor and Cognitive Outcomes in Nondisabled Low-Birth-Weight Adolescents: Early Determinants. Archives of Pediatrics & Adolescent Medicine 2006; 160: 1040-1046. DOI: 10.1001/archpedi.160.10.1040. 3. Smith CJ, Ryckman KK, Barnabei VM, et al. The impact of birth weight on cardiovascular disease risk in the Women's Health Initiative. Nutr Metab Cardiovasc Dis 2016; 26: 239-245. 2015/12/29. DOI: 10.1016/j.numecd.2015.10.015. 4. Jornayvaz FR, Vollenweider P, Bochud M, et al. Low birth weight leads to obesity, diabetes and increased leptin levels in adults: the CoLaus study. Cardiovasc Diabetol 2016; 15: 73. 2016/05/05. DOI: 10.1186/s12933-016-0389-2. 5. Schieve LA, Tian LH, Rankin K, et al. Population impact of preterm birth and low birth weight on developmental disabilities in US children. Ann Epidemiol 2016; 26: 267-274. 2016/03/22. DOI: 10.1016/j.annepidem.2016.02.012. 6. Mehra R, Keene DE, Kershaw TS, et al. Racial and ethnic disparities in adverse birth outcomes: Differences by racial residential segregation. SSM Popul Health 2019; 8: 100417-100417. DOI: 10.1016/j.ssmph.2019.100417. 7. Su D, Samson K, Hanson C, et al. Racial and ethnic disparities in birth Outcomes: A decomposition analysis of contributing factors. Prev Med Rep 2021; 23: 101456-101456. DOI: 10.1016/j.pmedr.2021.101456. 8. Burris HH, Lorch SA, Kirpalani H, et al. Racial disparities in preterm birth in USA: a biosensor of physical and social environmental exposures. Arch Dis Child 2019; 104: 931-935. 2019/03/08. DOI: 10.1136/archdischild-2018-316486. 9. Padula AM, Huang H, Baer RJ, et al. Environmental pollution and social factors as contributors to preterm birth in Fresno County. Environmental health : a global access science source 2018; 17: 70- 70. DOI: 10.1186/s12940-018-0414-x. 10. Huo W, Xia W, Wan Y, et al. Maternal urinary bisphenol A levels and infant low birth weight: A nested case-control study of the Health Baby Cohort in China. Environ Int 2015; 85: 96-103. 2015/09/19. DOI: 10.1016/j.envint.2015.09.005. 11. Li X, Huang S, Jiao A, et al. Association between ambient fine particulate matter and preterm birth or term low birth weight: An updated systematic review and meta-analysis. Environ Pollut 2017; 227: 596-605. 2017/05/02. DOI: 10.1016/j.envpol.2017.03.055. 12. Kojima H, Takeuchi S, Itoh T, et al. In vitro endocrine disruption potential of organophosphate flame retardants via human nuclear receptors. Toxicology 2013; 314: 76-83. 13. Trowbridge J, Gerona R, McMaster M, et al. Organophosphate and Organohalogen Flame- Retardant Exposure and Thyroid Hormone Disruption in a Cross-Sectional Study of Female Firefighters and Office Workers from San Francisco. Environmental Science & Technology 2022; 56: 440-450. DOI: 10.1021/acs.est.1c05140. 14. Yao Y, Li M, Pan L, et al. Exposure to organophosphate ester flame retardants and plasticizers during pregnancy: Thyroid endocrine disruption and mediation role of oxidative stress. Environ Int 2021; 146: 106215. 2020/10/29. DOI: 10.1016/j.envint.2020.106215. 15. Blum A, Behl M, Birnbaum L, et al. Organophosphate Ester Flame Retardants: Are They a Regrettable Substitution for Polybrominated Diphenyl Ethers? Environ Sci Technol Lett 2019; 6: 638- 649. 2020/06/05. DOI: 10.1021/acs.estlett.9b00582. 16. Li M, Yao Y, Wang Y, et al. Organophosphate ester flame retardants and plasticizers in a Chinese population: Significance of hydroxylated metabolites and implication for human exposure. Environ Pollut 2020; 257: 113633. 2019/11/26. DOI: 10.1016/j.envpol.2019.113633. 118 17. Patisaul HB, Behl M, Birnbaum LS, et al. Beyond Cholinesterase Inhibition: Developmental Neurotoxicity of Organophosphate Ester Flame Retardants and Plasticizers. Environ Health Perspect 2021; 129: 105001. 2021/10/07. DOI: 10.1289/ehp9285. 18. Yang J, Zhao Y, Li M, et al. A Review of a Class of Emerging Contaminants: The Classification, Distribution, Intensity of Consumption, Synthesis Routes, Environmental Effects and Expectation of Pollution Abatement to Organophosphate Flame Retardants (OPFRs). Int J Mol Sci 2019; 20 2019/06/20. DOI: 10.3390/ijms20122874. 19. Hou R, Xu Y and Wang Z. Review of OPFRs in animals and humans: Absorption, bioaccumulation, metabolism, and internal exposure research. Chemosphere 2016; 153: 78-90. DOI: https://doi.org/10.1016/j.chemosphere.2016.03.003. 20. Stapleton HM, Klosterhaus S, Eagle S, et al. Detection of organophosphate flame retardants in furniture foam and U.S. house dust. Environ Sci Technol 2009; 43: 7490-7495. 2009/10/24. DOI: 10.1021/es9014019. 21. Du J, Li H, Xu S, et al. A review of organophosphorus flame retardants (OPFRs): occurrence, bioaccumulation, toxicity, and organism exposure. Environ Sci Pollut Res Int 2019; 26: 22126-22136. 2019/06/28. DOI: 10.1007/s11356-019-05669-y. 22. He R, Li Y, Xiang P, et al. Organophosphorus flame retardants and phthalate esters in indoor dust from different microenvironments: Bioaccessibility and risk assessment. Chemosphere 2016; 150: 528- 535. 2015/11/21. DOI: 10.1016/j.chemosphere.2015.10.087. 23. Pantelaki I and Voutsa D. Organophosphate flame retardants (OPFRs): A review on analytical methods and occurrence in wastewater and aquatic environment. Sci Total Environ 2019; 649: 247-263. 2018/09/03. DOI: 10.1016/j.scitotenv.2018.08.286. 24. Thomas MB, Stapleton HM, Dills RL, et al. Demographic and dietary risk factors in relation to urinary metabolites of organophosphate flame retardants in toddlers. Chemosphere 2017; 185: 918-925. 2017/08/03. DOI: 10.1016/j.chemosphere.2017.07.015. 25. Wang X, Zhu Q, Yan X, et al. A review of organophosphate flame retardants and plasticizers in the environment: Analysis, occurrence and risk assessment. Sci Total Environ 2020; 731: 139071. 2020/05/22. DOI: 10.1016/j.scitotenv.2020.139071. 26. Wang Y, Li W, Martínez-Moral MP, et al. Metabolites of organophosphate esters in urine from the United States: Concentrations, temporal variability, and exposure assessment. Environ Int 2019; 122: 213-221. 2018/11/20. DOI: 10.1016/j.envint.2018.11.007. 27. Ospina M, Jayatilaka NK, Wong LY, et al. Exposure to organophosphate flame retardant chemicals in the U.S. general population: Data from the 2013-2014 National Health and Nutrition Examination Survey. Environ Int 2018; 110: 32-41. 2017/11/06. DOI: 10.1016/j.envint.2017.10.001. 28. Hoffman K, Daniels JL and Stapleton HM. Urinary metabolites of organophosphate flame retardants and their variability in pregnant women. Environ Int 2014; 63: 169-172. 2013/12/10. DOI: 10.1016/j.envint.2013.11.013. 29. Castorina R, Butt C, Stapleton HM, et al. Flame retardants and their metabolites in the homes and urine of pregnant women residing in California (the CHAMACOS cohort). Chemosphere 2017; 179: 159- 166. 2017/03/22. DOI: 10.1016/j.chemosphere.2017.03.076. 30. Doherty BT, Hoffman K, Keil AP, et al. Prenatal exposure to organophosphate esters and cognitive development in young children in the Pregnancy, Infection, and Nutrition Study. Environ Res 2019; 169: 33-40. 2018/11/09. DOI: 10.1016/j.envres.2018.10.033. 31. Kuiper JR, Stapleton HM, Wills-Karp M, et al. Predictors and reproducibility of urinary organophosphate ester metabolite concentrations during pregnancy and associations with birth outcomes in an urban population. Environ Health 2020; 19: 55. 2020/05/26. DOI: 10.1186/s12940-020-00610-0. 32. Hoffman K, Butt CM, Webster TF, et al. Temporal Trends in Exposure to Organophosphate Flame Retardants in the United States. Environmental Science & Technology Letters 2017; 4: 112-118. DOI: 10.1021/acs.estlett.6b00475. 119 33. Varshavsky JR, Robinson JF, Zhou Y, et al. Organophosphate Flame Retardants, Highly Fluorinated Chemicals, and Biomarkers of Placental Development and Disease During Mid-Gestation. Toxicol Sci 2021; 181: 215-228. 2021/03/08. DOI: 10.1093/toxsci/kfab028. 34. Ding J, Xu Z, Huang W, et al. Organophosphate ester flame retardants and plasticizers in human placenta in Eastern China. Sci Total Environ 2016; 554-555: 211-217. 2016/03/08. DOI: 10.1016/j.scitotenv.2016.02.171. 35. Zhao F, Chen M, Gao F, et al. Organophosphorus Flame Retardants in Pregnant Women and Their Transfer to Chorionic Villi. Environ Sci Technol 2017; 51: 6489-6497. 2017/05/19. DOI: 10.1021/acs.est.7b01122. 36. Bai XY, Lu SY, Xie L, et al. A pilot study of metabolites of organophosphorus flame retardants in paired maternal urine and amniotic fluid samples: potential exposure risks of tributyl phosphate to pregnant women. Environ Sci Process Impacts 2019; 21: 124-132. 2018/11/15. DOI: 10.1039/c8em00389k. 37. Li H, Su G, Zou M, et al. Effects of Tris(1,3-dichloro-2-propyl) Phosphate on Growth, Reproduction, and Gene Transcription of Daphnia magna at Environmentally Relevant Concentrations. Environmental Science & Technology 2015; 49: 12975-12983. DOI: 10.1021/acs.est.5b03294. 38. Farhat A, Crump D, Chiu S, et al. In ovo effects of two organophosphate flame retardants—TCPP and TDCPP—on pipping success, development, mRNA expression, and thyroid hormone levels in chicken embryos. toxicological sciences 2013; 134: 92-102. 39. Chen G, Jin Y, Wu Y, et al. Exposure of male mice to two kinds of organophosphate flame retardants (OPFRs) induced oxidative stress and endocrine disruption. Environ Toxicol Pharmacol 2015; 40: 310-318. 2015/07/18. DOI: 10.1016/j.etap.2015.06.021. 40. Yu L, Jia Y, Su G, et al. Parental transfer of tris(1,3-dichloro-2-propyl) phosphate and transgenerational inhibition of growth of zebrafish exposed to environmentally relevant concentrations. Environ Pollut 2017; 220: 196-203. 2016/09/21. DOI: 10.1016/j.envpol.2016.09.039. 41. Sun L, Tan H, Peng T, et al. Developmental neurotoxicity of organophosphate flame retardants in early life stages of Japanese medaka (Oryzias latipes). Environ Toxicol Chem 2016; 35: 2931-2940. 2016/05/06. DOI: 10.1002/etc.3477. 42. Zhu Y, Ma X, Su G, et al. Environmentally Relevant Concentrations of the Flame Retardant Tris(1,3-dichloro-2-propyl) Phosphate Inhibit Growth of Female Zebrafish and Decrease Fecundity. Environ Sci Technol 2015; 49: 14579-14587. 2015/10/30. DOI: 10.1021/acs.est.5b03849. 43. Wang Q, Liang K, Liu J, et al. Exposure of zebrafish embryos/larvae to TDCPP alters concentrations of thyroid hormones and transcriptions of genes involved in the hypothalamic–pituitary– thyroid axis. Aquatic toxicology 2013; 126: 207-213. 44. Zhong X, Wu J, Ke W, et al. Neonatal exposure to organophosphorus flame retardant TDCPP elicits neurotoxicity in mouse hippocampus via microglia-mediated inflammation in vivo and in vitro. Arch Toxicol 2020; 94: 541-552. 2020/01/03. DOI: 10.1007/s00204-019-02635-y. 45. Zhong X, Yu Y, Wang C, et al. Hippocampal proteomic analysis reveals the disturbance of synaptogenesis and neurotransmission induced by developmental exposure to organophosphate flame retardant triphenyl phosphate. J Hazard Mater 2021; 404: 124111. 2020/11/15. DOI: 10.1016/j.jhazmat.2020.124111. 46. Ferrante J. Toxicity review of aromatic phosphate plasticizers. Bethesda, MD 1999; 20814. 47. Dishaw LV, Powers CM, Ryde IT, et al. Is the PentaBDE replacement, tris (1,3-dichloro-2- propyl) phosphate (TDCPP), a developmental neurotoxicant? Studies in PC12 cells. Toxicol Appl Pharmacol 2011; 256: 281-289. 2011/01/19. DOI: 10.1016/j.taap.2011.01.005. 48. Fu J, Han J, Zhou B, et al. Toxicogenomic responses of zebrafish embryos/larvae to tris (1, 3- dichloro-2-propyl) phosphate (TDCPP) reveal possible molecular mechanisms of developmental toxicity. Environmental science & technology 2013; 47: 10574-10582. 49. Shi Q, Tsui MMP, Hu C, et al. Acute exposure to triphenyl phosphate (TPhP) disturbs ocular development and muscular organization in zebrafish larvae. Ecotoxicol Environ Saf 2019; 179: 119-126. 2019/04/30. DOI: 10.1016/j.ecoenv.2019.04.056. 120 50. Chen H, Wang P, Du Z, et al. Oxidative stress, cell cycle arrest, DNA damage and apoptosis in adult zebrafish (Danio rerio) induced by tris(1,3-dichloro-2-propyl) phosphate. Aquat Toxicol 2018; 194: 37-45. 2017/11/18. DOI: 10.1016/j.aquatox.2017.11.001. 51. Rock KD, St Armour G, Horman B, et al. Effects of Prenatal Exposure to a Mixture of Organophosphate Flame Retardants on Placental Gene Expression and Serotonergic Innervation in the Fetal Rat Brain. Toxicol Sci 2020; 176: 203-223. 2020/04/04. DOI: 10.1093/toxsci/kfaa046. 52. Hoffman K, Stapleton HM, Lorenzo A, et al. Prenatal exposure to organophosphates and associations with birthweight and gestational length. Environ Int 2018; 116: 248-254. 2018/04/27. DOI: 10.1016/j.envint.2018.04.016. 53. Luo D, Liu W, Tao Y, et al. Prenatal Exposure to Organophosphate Flame Retardants and the Risk of Low Birth Weight: A Nested Case-Control Study in China. Environ Sci Technol 2020; 54: 3375- 3385. 2020/02/29. DOI: 10.1021/acs.est.9b06026. 54. Bastain TM, Chavez T, Habre R, et al. Study Design, Protocol and Profile of the Maternal And Developmental Risks from Environmental and Social Stressors (MADRES) Pregnancy Cohort: a Prospective Cohort Study in Predominantly Low-Income Hispanic Women in Urban Los Angeles. BMC Pregnancy Childbirth 2019; 19: 189-189. DOI: 10.1186/s12884-019-2330-7. 55. Hornung RW and Reed LD. Estimation of Average Concentration in the Presence of Nondetectable Values. Applied Occupational and Environmental Hygiene 1990; 5: 46-51. DOI: 10.1080/1047322X.1990.10389587. 56. Romano ME, Hawley NL, Eliot M, et al. Variability and predictors of urinary concentrations of organophosphate flame retardant metabolites among pregnant women in Rhode Island. Environmental Health 2017; 16: 1-11. 57. Dunlop AL, Essalmi AG, Alvalos L, et al. Racial and geographic variation in effects of maternal education and neighborhood-level measures of socioeconomic status on gestational age at birth: Findings from the ECHO cohorts. PLoS One 2021; 16: e0245064. 2021/01/09. DOI: 10.1371/journal.pone.0245064. 58. Lourenço BH, Lima DL, Vivanco E, et al. Agreement between antenatal gestational age by ultrasound and clinical records at birth: A prospective cohort in the Brazilian Amazon. PLoS One 2020; 15: e0236055. 2020/07/15. DOI: 10.1371/journal.pone.0236055. 59. Obstetricians ACo and Gynecologists. Methods for estimating the due date. Committee Opinion No. 700. Obstet Gynecol 2017; 129: e150-e154. 60. Aris IM, Kleinman KP, Belfort MB, et al. A 2017 US Reference for Singleton Birth Weight Percentiles Using Obstetric Estimates of Gestation. Pediatrics 2019; 144 2019/06/16. DOI: 10.1542/peds.2019-0076. 61. Textor J, van der Zander B, Gilthorpe MS, et al. Robust causal inference using directed acyclic graphs: the R package 'dagitty'. Int J Epidemiol 2016; 45: 1887-1894. 2017/01/17. DOI: 10.1093/ije/dyw341. 62. Howe CG, Claus Henn B, Eckel SP, et al. Prenatal Metal Mixtures and Birth Weight for Gestational Age in a Predominately Lower-Income Hispanic Pregnancy Cohort in Los Angeles. Environ Health Perspect 2020; 128: 117001. 2020/11/04. DOI: 10.1289/ehp7201. 63. Howe CG, Claus Henn B, Farzan SF, et al. Prenatal metal mixtures and fetal size in mid- pregnancy in the MADRES study. Environmental Research 2021; 196: 110388. DOI: https://doi.org/10.1016/j.envres.2020.110388. 64. Liu W, Luo D, Xia W, et al. Prenatal exposure to halogenated, aryl, and alkyl organophosphate esters and child neurodevelopment at two years of age. Journal of Hazardous Materials 2021; 408: 124856. DOI: https://doi.org/10.1016/j.jhazmat.2020.124856. 65. Bobb JF, Valeri L, Claus Henn B, et al. Bayesian kernel machine regression for estimating the health effects of multi-pollutant mixtures. Biostatistics 2015; 16: 493-508. 2014/12/24. DOI: 10.1093/biostatistics/kxu058. 121 66. Bobb JF, Claus Henn B, Valeri L, et al. Statistical software for analyzing the health effects of multiple concurrent exposures via Bayesian kernel machine regression. Environmental Health 2018; 17: 67. DOI: 10.1186/s12940-018-0413-y. 67. Crawford KA, Hawley N, Calafat AM, et al. Maternal urinary concentrations of organophosphate ester metabolites: associations with gestational weight gain, early life anthropometry, and infant eating behaviors among mothers-infant pairs in Rhode Island. Environmental Health 2020; 19: 97. DOI: 10.1186/s12940-020-00648-0. 68. Yang W, Braun JM, Vuong AM, et al. Associations of gestational exposure to organophosphate esters with gestational age and neonatal anthropometric measures: The HOME study. Environ Pollut 2023; 316: 120516. 2022/11/08. DOI: 10.1016/j.envpol.2022.120516. 69. Doherty BT, Hammel SC, Daniels JL, et al. Organophosphate Esters: Are These Flame Retardants and Plasticizers Affecting Children's Health? Curr Environ Health Rep 2019; 6: 201-213. 2019/11/23. DOI: 10.1007/s40572-019-00258-0. 70. Xu T, Wang Q, Shi Q, et al. Bioconcentration, metabolism and alterations of thyroid hormones of Tris (1, 3-dichloro-2-propyl) phosphate (TDCPP) in Zebrafish. Environmental toxicology and pharmacology 2015; 40: 581-586. 71. Williams GR. Neurodevelopmental and neurophysiological actions of thyroid hormone. J Neuroendocrinol 2008; 20: 784-794. 2008/07/08. DOI: 10.1111/j.1365-2826.2008.01733.x. 72. Percy Z, Vuong AM, Xu Y, et al. Maternal Urinary Organophosphate Esters and Alterations in Maternal and Neonatal Thyroid Hormones. Am J Epidemiol 2021; 190: 1793-1802. 2021/03/30. DOI: 10.1093/aje/kwab086. 73. Liu X, Ji K, Jo A, et al. Effects of TDCPP or TPP on gene transcriptions and hormones of HPG axis, and their consequences on reproduction in adult zebrafish (Danio rerio). Aquatic toxicology 2013; 134: 104-111. 74. Liu X, Ji K and Choi K. Endocrine disruption potentials of organophosphate flame retardants and related mechanisms in H295R and MVLN cell lines and in zebrafish. Aquatic toxicology 2012; 114: 173- 181. 75. Meeker JD and Stapleton HM. House Dust Concentrations of Organophosphate Flame Retardants in Relation to Hormone Levels and Semen Quality Parameters. Environmental Health Perspectives 2010; 118: 318-323. DOI: doi:10.1289/ehp.0901332. 76. Carignan CC, Mínguez-Alarcón L, Williams PL, et al. Paternal urinary concentrations of organophosphate flame retardant metabolites, fertility measures, and pregnancy outcomes among couples undergoing in vitro fertilization. Environ Int 2018; 111: 232-238. 2017/12/15. DOI: 10.1016/j.envint.2017.12.005. 77. Kojima H, Takeuchi S, Van den Eede N, et al. Effects of primary metabolites of organophosphate flame retardants on transcriptional activity via human nuclear receptors. Toxicology Letters 2016; 245: 31-39. DOI: https://doi.org/10.1016/j.toxlet.2016.01.004. 78. Liu X, Jung D, Jo A, et al. Long-term exposure to triphenylphosphate alters hormone balance and HPG, HPI, and HPT gene expression in zebrafish (Danio rerio). Environ Toxicol Chem 2016; 35: 2288- 2296. 2016/02/13. DOI: 10.1002/etc.3395. 79. Ma Z, Tang S, Su G, et al. Effects of tris (2-butoxyethyl) phosphate (TBOEP) on endocrine axes during development of early life stages of zebrafish (Danio rerio). Chemosphere 2016; 144: 1920-1927. 2015/11/08. DOI: 10.1016/j.chemosphere.2015.10.049. 80. Davis LK, Pierce AL, Hiramatsu N, et al. Gender-specific expression of multiple estrogen receptors, growth hormone receptors, insulin-like growth factors and vitellogenins, and effects of 17β- estradiol in the male tilapia (Oreochromis mossambicus). General and Comparative Endocrinology 2008; 156: 544-551. DOI: https://doi.org/10.1016/j.ygcen.2008.03.002. 81. Guan R, Li N, Wang W, et al. The adverse outcome pathway (AOP) of estrogen interference effect induced by triphenyl phosphate (TPP): Integrated multi-omics and molecular dynamics approaches. Ecotoxicology and Environmental Safety 2022; 234: 113387. DOI: https://doi.org/10.1016/j.ecoenv.2022.113387. 122 82. Brown ZA, Schalekamp-Timmermans S, Tiemeier HW, et al. Fetal sex specific differences in human placentation: A prospective cohort study. Placenta 2014; 35: 359-364. DOI: https://doi.org/10.1016/j.placenta.2014.03.014. 83. Leonetti C, Butt CM, Hoffman K, et al. Brominated flame retardants in placental tissues: associations with infant sex and thyroid hormone endpoints. Environmental Health 2016; 15: 113. DOI: 10.1186/s12940-016-0199-8. 84. Lapehn S and Paquette AG. The Placental Epigenome as a Molecular Link Between Prenatal Exposures and Fetal Health Outcomes Through the DOHaD Hypothesis. Curr Environ Health Rep 2022; 9: 490-501. 2022/04/30. DOI: 10.1007/s40572-022-00354-8. 85. Lynn R, Wong K, Garvie-Gould C, et al. Disposition of the flame retardant, tris (1, 3-dichloro-2- propyl) phosphate, in the rat. Drug metabolism and Disposition 1981; 9: 434-441. 86. Nomeir A, Kato S and Matthews H. The metabolism and disposition of tris (1, 3-dichloro-2- propyl) phosphate (Fyrol FR-2) in the rat. Toxicol Appl Pharmacol 1981; 57: 401-413. 87. MINEGISHI K, KUREBAYASHI H, NAMBARU S, et al. Comparative studies on absorption, distribution, and excretion of flame retardants halogenated alkyl phosphate in rats. Eisei kagaku 1988; 34: 102-114. 88. Sasaki K, Takeda M and Uchiyama M. Toxicity, absorption and elimination of phosphoric acid triesters by killifish and goldfish. Bull Environ Contam Toxicol 1981; 27: 775-782. 123 CHAPTER 4 STUDY 2: Prenatal Exposures to Organophosphate Ester Metabolites and Early Motor Development in the MADRES Cohort ABSTRACT Background: Organophosphate esters (OPEs) are used as flame retardants and plasticizers on a variety of consumer products. OPEs are increasingly considered neurotoxicants and may impact the development of early gross and fine motor skills. We evaluated overall and sex-specific associations between prenatal OPE exposures and infant motor development. Methods: We measured third trimester urinary concentrations of nine OPE metabolites (diphenyl phosphate (DPHP), dibutyl phosphate + di-isobutyl phosphate (DNBP+DIBP), bis(1,3,-dichloro-2-propyl) phosphate (BDCIPP), bis(2-chloroethyl) phosphate (BCEP), bis(butoxethyl) phosphate (BBOEP), bis(1- chloro-2-propyl) phosphate (BCIPP), bis(2-ethylhexyl) phosphate (BEHP), bis(2-methylphenyl) phosphate (BMPP), and dipropyl phosphate (DPRP)) in 329 mother-infant dyads participating in the MADRES cohort. Child gross and fine motor development at 6, 9, 12, and 18-months were assessed with the Ages and Stages Questionnaire-3 (ASQ-3) and operationalized in models using both dichotomous instrument-specific cutoffs (above/below) for possible motor developmental delay and a reverse scored and treated as an integer count version of the ASQ-3 continuous scores. Urinary OPE metabolites with >60% detection were modeled continuously after specific gravity adjustment and natural log-transformation, while metabolites with <60% detection were modeled dichotomously (detected/not-detected). We fit both mixed effects logistic regression and negative binomial mixed models relating each OPE metabolite to fine and gross motor development outcomes. Sex-specific effects for each metabolite and time at ASQ-3 administration were assessed using a statistical interaction term and in fully sex-stratified models. Models were adjusted for key demographic and study design covariates. Results: Thirty-one percent of children had gross motor and 23% had fine motor scores above the ASQ-3 at-risk cutoffs at least once across infancy. A doubling in prenatal DPHP exposure was associated with 26% 124 increased odds of being at risk for fine motor delays (OR=1.26, 95% CI: 1.02, 1.57) and 11% decrease in expected fine motor scores (IRR=1.11, 95% CI: 1.01, 1.22). Detectable levels of BMPP were associated with increased gross motor performance (IRR: 0.77, 95% CI: 0.63, 0.94) when compared to participants with non-detectable levels of BMPP; however, associations were not significant across clinically relevant cut-offs. We also observed significant interactions by infant sex for associations of DPRP with gross motor development (p<0.05) and BCIPP with fine motor development (p=0.02), with female infants showing greater odds of being at risk of motor delays for both DPRP (males vs females OR (95% CI)= 0.27 (0.44, 3.51) vs 1.48 (0.71, 3.09)) and BCIPP (males vs females OR (95% CI)= 0.76 (0.31, 1.90) vs 2.72 (1.27, 5.85)). Conclusion: We found evidence of adverse effects of prenatal OPE metabolites on infant fine motor scores and mixed effects on gross motor development, with more adverse effects among female infants, suggesting sex-specific impacts of these metabolites on neuromotor development. INTRODUCTION Infant motor development is an important neurodevelopmental milestone tied to various childhood outcomes, including later childhood motor development, cognitive and social development, and metabolic health. 1-4 Early motor development begins as early as the embryonic period, with the establishment of the primitive patterning of the sensorimotor regions, and continues to develop until adulthood. 5,6 Given the rapid cascade of neurological processes that occur during the fetal period, in-utero motor development is particularly vulnerable and susceptible to environmental insults. 7 For instance, neurotoxic environmental exposures during the in-utero period may result in alterations to neuronal migration and disruptions to the apoptosis cascade which may adversely impact motor development via the sensorimotor system. 7-11 Since environmental exposures are modifiable risk factors, their impacts on early motor development are important to understand. 12 Early exposures to other organophosphate chemicals, specifically pesticides, have been found to adversely impact early motor development. 13,14 The impacts of other similar organophosphate chemicals on early motor development, such as organophosphate esters (OPEs) used as flame retardants and plasticizers, 125 may have similar adverse impacts on motor development via non-cholinergic pathways, according to emerging literature. 15,16 OPEs are a class of organic, anthropogenic chemicals applied on various industrial, consumer, and electronic products as plasticizers, lubricants, and as chemical flame retardants to meet flammability regulations. 17,18 Due to the phase of out of polybrominated diphenyl ethers (PBDEs) over health and bioaccumulation related concerns, OPEs have dramatically increased in use. 18,19 Since OPEs are physically bound with a product matrix rather than chemically bound, they may easily volatilize and leach into surrounding environments including soil, surface water, sediment, and indoor dust particles, facilitating exposures to OPEs among the general population. 20-24 Biomonitoring studies indicate high detectable frequencies of OPE metabolites in the general population (>95% detect frequency), particularly of diphenyl phosphate (DPHP; parent compound= triphenyl phosphate (TPHP)) and bis(1,3-dichloro-2-propyl) phosphate (BDCIPP; parent compound= tris(1,3-dichloropropyl) (TDCIPP)). 25-29 Common OPE exposure routes include inhalation, dermal contact, and ingestion of air and dust particles containing OPEs as well as dietary ingestion of OPEs in food and drinking water. 21,30-32 After exposure, OPEs metabolize into their respective mono- or diesters which are then primarily excreted in urine. 20,33-38 Coupled with their structural similarity to other neurotoxic chemicals, ubiquitous exposure, and fetal neurological susceptibility, OPEs may pose a significant risk to early motor development. 7,15,17 The objective of this study was to evaluate the impacts of prenatal OPE metabolites on infant motor development throughout infancy (6, 9, 12, and 18 months of age) among 329 mother-infant dyads participating in the MADRES cohort study. Given previous literature suggesting possible sex-specific biological impacts, 28,39,40 this study additionally evaluated whether infant sex modified the association between OPE metabolites and early motor development. METHODS Study Design The MADRES study is an ongoing prospective pregnancy cohort of predominately low-income Hispanic/Latina women living in urban Los Angeles. A detailed description of the MADRES study protocol 126 and population have been previously described elsewhere. 41 Briefly, participants were recruited into the study prior to 30 weeks’ gestation at three partner community health clinics, one private obstetrics and gynecology practice in Los Angeles, and through self-referrals from community meetings and local advertisements. Participants were eligible at time of recruitment if they were: (1) less than 30 weeks’ gestation, (2) over 18 years of age, and (3) fluent in English or Spanish. Participants were excluded if they had: (1) multiple gestation, (2) a physical, mental, or cognitive disability that prevented participation or ability to provide consent, (3) current incarceration, and (4) HIV positive status. Informed consent was obtained at study entry for each participant and the study was approved by the University of Southern California’s Institutional Review Board. Urinary OPE metabolite concentrations were measured in 426 MADRES participants’ samples collected during the third trimester study visit (mean gestational age at collection: 31.53 weeks). Participants included in this analysis were recruited from March 2016 to October 2019. As shown in the consort diagram (Figure 1), there were a total of 329 mother-infant dyads analyzed in this study with complete data on prenatal metabolite concentrations, motor development outcomes from at least one timepoint, and information on key covariates. The 97 participant dyads excluded from this analytical sample had similar key demographic characteristics as the 329 analyzed in this study (Table S1). OPE Metabolite Analysis Maternal single spot urine samples were collected in 90 mL sterile specimen containers. Urine specimens were separated into 1.5 mL aliquot cryovials and stored at -80º Celsius prior to shipment. Specific gravity was measured in urine samples using a digital handheld refractometer (ATAGO PAL-10s pocket refractometer) and samples were sent to the Wadsworth Center Human Health Exposure Analysis Resource (HHEAR) for analysis of nine OPE metabolites listed in Chapter 1 (Table 1): diphenyl phosphate (DPHP), sum of dibutyl phosphate (DNBP) and di-isobutyl phosphate (DIBP), bis(1,3,-dichloro-2-propyl) phosphate (BDCIPP), bis(2-chloroethyl) phosphate (BCEP), bis(butoxethyl) phosphate (BBOEP), bis(1- chloro-2-propyl) phosphate (BCIPP), bis(2-ethylhexyl) phosphate (BEHP), bis(2-methylphenyl) phosphate (BMPP), dipropyl phosphate (DPRP). 127 A detailed description of methods used for OPEs analysis has been detailed in Chapter 2 (Prenatal Measures of Urinary OPE Metabolites) and previously described. 32 Briefly, high-performance liquid chromatography (HPLC, ExionLC™ system; SCIEX, Redwood City, CA, USA), coupled with an AB SCIEX QTRAP 5500+ triple quadrupole mass spectrometer (Applied Biosystems, Foster City, CA, USA), was used in the identification and quantification of target compounds. Target analytes limit of detection (LOD) ranged from 0.012 to 0.044 ng/mL. As a result of the poor chromatographic separation and co- elution of peaks accompanying a similar mass transition for DNBP and DIBP in the lab’s mass spectrometer, a summed concentration for the metabolites of dibutyl phosphate and di-isobutyl phosphate (DNBP + DIBP) was evaluated. OPE metabolites with concentrations below the LOD were imputed using the LOD/√2 and then specific gravity adjusted using the following formula: Pc=P[(SGm-1)/(SG-1)], where Pc is the specific gravity corrected toxicant concentration (ng/mL), P is the observed toxicant concentration (ng/mL), SGm is the median SG value among the study population (median=1.016), and SG=the SG value of the sample. 42 Health Outcome Assessment The ASQ-3 is a validated screening instrument which contains five developmental subscales examining neurodevelopmental performance across gross motor, fine motor, problem solving, communication, and personal-social domains for use among children 3 to 61 months old. 43 At the 6-month study timepoint, only the gross and fine motor subscales were administered out of concern for participant burden. Although all developmental domains were administered at the 9-, 12-, and 18-months study timepoints, only the motor subscales were assessed in this analysis to leverage available longitudinal data beginning earlier in infancy. ASQ-3 administration at the four study timepoints occurred when the child turned the age specified at the study visit timepoint (6, 9, 12, 18 months) and continued until the close of that study visit window, with the following ranges in windows by study visit timepoints: (1) 6 months + 6 weeks, (2) 9 months + 6 weeks, (3) 11 months + 8 weeks, and (4) 18 months + 6 weeks, respectively. 128 ASQ-3 motor subscales were administered and scored according to the ASQ-3 protocol. 43 Scores range from 0 to 60, with higher scores indicating better performance in the respective age specific developmental domain. Predetermined cut-off scores for further reassessment with a healthcare professional are scores that are at least 2 standard deviations lower from the average ASQ-3 scores of a normative US sample, while a monitoring zone marking potential areas of future concern is designated as being at least 1 but less than 2 standard deviations lower from the ASQ-3 scores of the average normative US sample. Covariates Study covariates were identified a priori based on previous literature evaluating impacts of neurotoxic chemicals on neurodevelopment and visualized using a Directed Acyclic Graph (DAG) created using DAGitty (Figure S1). 24,28,40,44-46 Models were adjusted for study design variables, such as recruitment site, gestational age at urine sample collection, season of sample collection, and infant adjusted age at questionnaire administration and variables identified in the minimal sufficient adjustment set, included maternal age, parity, pre-pregnancy BMI, race/ethnicity, income, and education. Exceptions to these criteria included: 1) adjustment for infant sex in overall models given the importance of infant sex as a predictor of neurodevelopment outcomes, and 2) exclusion of prenatal smoking status as an adjustment variable due to the small frequency of reported maternal smoking (n=7, 2.1%). Sensitivity analyses removing participants who smoked during pregnancy were performed instead. Participants’ demographic information, including maternal age (years), household annual income during pregnancy (<$30,000, ≥ $30,000, do not know), education (≤12 th grade, >12 th grade), race/ethnicity (White non-Hispanic, Black non-Hispanic, Hispanic, Multiracial non-Hispanic/Other non-Hispanic), maternal smoking during pregnancy (yes, no), and parity (first born, ≥ second born, missing), were collected via interviewer administered questionnaires in the participant’s preferred language (English or Spanish). Pre-pregnancy BMI was calculated using participant-reported pre-pregnancy weight and standing height measured by study staff in the first study visit using a commercial stadiometer (Perspectives Enterprises model P-AIM-101). Infant sex assigned at birth was primarily abstracted from electronic medical records 129 (n=318, 96.7%), followed by maternal-reported child sex (n=11, 3.3%) for cases in which abstracted sex was unavailable. Statistical Analysis Descriptive statistics, including participant characteristics and OPE metabolite distributions, including percentiles, medians, geometric means, and detect frequencies, were assessed. Since OPE metabolites are relatively right skewed, Spearman correlations were used to evaluate the degree of association between metabolites and associations between OPE metabolites and categorical covariates were assessed using Kruskal Wallis and Pearson Chi-Square tests. Both clinically relevant cut-offs of the ASQ-3 and continuous ASQ-3 scores were used to examine possible risk of developmental delays and performance across motor scales at 6-, 9-, 12-, and 18-months, using generalized linear mixed effects regression models (logistic and negative binomial, respectively) with participant-level random intercepts. ASQ-3 gross and fine motor scores were dichotomized for logistic mixed models by combining the “suggest monitoring” and “suggest further assessment with a health care professional” cutoffs into a single “higher risk of motor delay” category (≥1 standard deviations lower from the average ASQ-3 scores of a normative US sample) to compare to typically developing infants (reference). Since ASQ-3 motor scores at all timepoints were left skewed, ASQ-3 motor scores were operationalized in negative binomial mixed models using a previously described reverse scored and treated as an integer count version of the ASQ-3, which used the natural 5- to 10-point increment increase of the scores to reverse score and convert each 5-point increment into a single count. 47 According to the scoring manual, ASQ-3 scores are adjusted for missing items which can cause deviation from the natural 5- to 10- point incremental scores and non-integer values. In such cases, we rounded to the nearest integer, impacting a small number of observations for both gross (n=2, 0.3%) and fine (n=5, 0.8%) motor scores. OPE metabolites with a detect frequency >60% were natural log transformed and analyzed continuously, and beta estimates were then transformed to a doubling in exposure to facilitate interpretation. OPE metabolites with a detect frequency <60% were analyzed as detect versus non-detect groups. In an attempt to understand the cumulative impacts of OPEs exposures on motor development during infancy, a 130 total molar sum of the 9 OPE concentrations was calculated (∑OPEs nmol/L) using methods previously described in the literature and assessed across models. 40 We additionally evaluated whether infant sex modified the association between OPEs and motor development in infancy by including a statistical interaction term in models and also stratifying models by infant sex, given prior evidence indicating potentially sex-specific effects from OPEs exposures. To assess possible changes in the association between OPEs and risk of motor delays across time in mixed effects logistic regression models and possible changes in scores across time in negative binomial mixed models, we also included a statistical interaction term by time of ASQ-3 administration. We additionally explored possible non-linear associations between OPE metabolites with a detect frequency >60% (n=5) and gross and fine motor development using generalized additive mixed models (GAMMs) for logistic regressions and negative binomial regressions, with a smoothing term for OPE metabolites and participant-level random intercepts. Several sensitivity analyses were performed to assess the impact of prenatal OPEs on motor development across administration timepoints and to evaluate the robustness of our results. For one, timepoint specific associations were assessed for each timepoint of the ASQ-3 with sufficient sample size for model convergence (6- and 12-month timepoints). Additionally, models excluding maternal participants who smoked during pregnancy and models removing infants born at <37 weeks’ gestation at birth were also performed. We further assessed the impact of rounding non-integers to the nearest count in negative binomial models by rounding all integers up and then rounding all integers down. All data was managed using SAS Version 9.4. Generalized linear mixed models were estimated using the “glmmTMB” package in R (v 4.3.1) and GAMMs were estimated using the “mgcv” package. All necessary model assumptions were met and the significance level for all models was set at an alpha of 0.05. 131 RESULTS Descriptive Statistics As shown in Table 1, maternal participants were an average of 29.2±6.1 years old at study recruitment and were predominately Hispanic (81.2%). About half of participants completed at most high school (56.5%) and had an annual household income of less than $30,000 (46.5%). Infants were 51.7% female, primarily second born or more (62.9%), and were born at an average gestational age of 39.1±1.5 weeks. Prenatal OPE Metabolite Concentrations OPE metabolite distributions are shown in Table 2. Detect frequencies were greater than 60% for five metabolites (DPHP, DNBP+DIBP, BDCIPP, BCEP, BBOEP) and less than 60% for four metabolites (BCIPP, BEHP, BMPP, DPRP). The highest OPE median concentrations were seen for BDCIPP (1.32 ng/mL) and DPHP (0.78 ng/mL). Figure 2 illustrates Spearman correlations between all OPE metabolites analyzed, with the highest correlation between BDCIPP and DPHP (𝜌 = 0.26, p<0.01). Motor Development Assessments The distribution of all ASQ-3 scores (on the original scale), as well as the frequency of gross and fine motor scores below the clinically recommended cut-offs, are shown in Figure 3. The percent of participants with ASQ-3 scores below the recommended cut-offs for monitoring and further assessment (≥1 standard deviations from the average ASQ-3 scores of a normative sample) was 20.3% for the 6-month gross motor scale, 14.8% for the fine motor scale, 27.4% for the 9-month gross motor scale, 17.8% for the fine motor scale, 24.9% for the 12-month gross motor scale, 13.5% for the fine motor scale, 12.8% for the 18-month gross motor scale, and 18.0% for the 18-month fine motor scale. Thirty-one percent and 23% of children had gross and fine motor scores below the ASQ-3 at-risk cutoffs at least once across the four timepoints. Additionally, of the 31% and 23% of children who had gross and fine motor scores below the at-risk cutoffs at least once, 23% scored below the gross motor ASQ-3 at risk cut-off at least twice or more and 16% scored below the fine motor ASQ-3 cut-off at least twice or more across the four timepoints. 132 Prenatal OPE Metabolites and Motor Development Across Infancy As shown in Table 3, we found that a doubling in prenatal DPHP exposure was associated with 26% (OR: 1.26, 95% CI: 1.02, 1.57) increased odds of being at risk of fine motor delays and an 11% (IRR: 1.11, 95% CI: 1.01, 1.22) decrease in expected fine motor scores among children, after adjusting for key covariates (Table 3). Detectable levels of prenatal BMPP were associated with higher expected gross motor performance in children (IRR: 0.77, 95% CI: 0.63, 0.94) when compared to maternal participants with no detectable levels of BMPP during pregnancy; however, participants with detected levels of prenatal BMPP did not have babies with significantly decreased odds of being at risk for gross motor delays (OR: 0.60, 95% CI: 0.33, 1.11). We did not find any other significant associations between prenatal OPE metabolite concentrations and gross and fine motor development across infancy. Although non-significant, there was a pattern between DNBP+DIBP and 30% (OR: 1.30, 95% CI: 0.96, 1.74) increased odds of being at risk of gross motor delays and 8% (IRR: 1.08, 95% CI: 0.97, 1.19) decrease in expected gross motor scores among children. When we evaluated sex-specific associations between prenatal OPE metabolites on children’s gross and fine motor development across infancy, we found a statistically significant interaction between prenatal DPRP concentrations and infant sex on children’s gross motor development in both mixed effects logistic regression models (p<0.05) and negative binomial mixed models (p<0.01). Within stratified models, female infants showed patterns of greater odds of being at risk for gross motor delays (OR: 1.48, 95% CI: 0.71, 3.01) and lower expected performance on the gross motor scale (IRR: 1.29, 95% CI: 0.94, 1.77) with detectable prenatal levels of DPRP, relative to those with non-detectable prenatal levels of DPRP. Male infants showed protective patterns of gross motor effects from detectable prenatal levels of DPRP (OR=0.27, 95% CI: 0.06, 1.29) and an association with improved performance on the gross motor scale (IRR= 0.63, 95% CI: 0.47, 0.86), when compared to those with non-detectable levels of DPRP. We also found a statistically significant interaction between prenatal BCIPP and infant sex on children’s fine motor development in the mixed effects logistic regression (p=0.02). Again, female infants with detectable levels of prenatal BCIPP had greater odds of being at risk for fine motor delays (OR: 2.72, 95% CI: 1.27, 5.85) 133 relative to those without detectable levels of prenatal BCIPP, while male infants had patterns suggestive of protective effects (OR: 0.76, 95% CI: 0.31, 1.90). We also found a statistically significant interaction between prenatal BCEP and increased risk of possible gross motor delay with time (p=0.04), with an estimated 0.9% decrease in possible risk of gross motor delays per infant adjusted age in weeks at ASQ-3 administration. We did not find any other significant interactions between prenatal OPEs and timepoint of ASQ-3 administration on gross and fine motor development across models. We found no statistically significant associations between the molar sum of OPEs and gross and fine motor scores. However, patterns between the molar sum of OPEs and motor scores were generally suggestive of increased risk of possible gross and fine motor delays and decreased gross and fine motor performance. Longitudinal associations between prenatal OPEs and gross and fine motor development were generally linear in GAMM models (Figure S2 and Figure S3) and consistent with the results observed in the mixed-effects logistic and negative binomial mixed models. For instance, prenatal DPHP exposure was associated with significantly increased risk of fine motor delays (p=0.04) and poorer fine motor skills in children (p=0.03). In sensitivity analyses evaluating timepoint-specific associations between prenatal OPE metabolites and motor development in logistic regression (Table S2) and negative binomial regression models (Table S3), we observed 30% increased odds of being at risk of fine motor delays (OR=1.30, 95% CI: 1.01, 1.67) and 14% decrease in expected fine motor scores at 6-months (IRR: 1.14, 95% CI: 1.01, 1.29). Sensitivity analyses excluding maternal participants who smoked during pregnancy (Table S4) were generally consistent with the main results. However, in sensitivity analyses removing infants born preterm (<37 weeks’ gestation (n=28)), the associations were stronger between prenatal DPHP and children’s fine motor development across both modeling frameworks, but the protective association between prenatal BMPP and gross motor development was attenuated and no longer significant (Table S5). Negative binomial regression models with non-integers rounded up and down (Table S6) had consistent estimates to those observed in our main results. 134 DISCUSSION We found evidence of adverse impacts from prenatal exposures to DPHP on fine motor development among the MADRES cohort. Evidence of sex-specific impacts from prenatal OPE exposures on gross and fine motor development was also observed, with adverse associations between DPRP and children’s gross motor development and BCIPP and children’s fine motor development among female infants only. We also found an increase in gross motor performance from detectable levels of prenatal BMPP that no longer remained when we removed preterm infants. Overall, our results are suggestive of adverse impacts from prenatal OPE metabolite concentrations on fine motor development during early infancy and stress the importance of considering the sex-specific impacts of prenatal OPEs on early neurodevelopment. It is additionally important to note that since neurodevelopmental milestones are intrinsically linked to one another, particularly during infancy, any disruptions to infants’ motor development could disrupt their cognitive, social, or language development, suggesting potentially far reaching neurodevelopmental impacts from prenatal OPE exposures which should be further explored in future research. 48-50 Growing epidemiological evidence has found associations between prenatal OPE metabolite concentrations and adverse neurodevelopmental outcomes across various domains, including cognitive and fine motor development; however, results have been mixed. Our results are somewhat consistent with a study by Doherty et al. among 227 participants from the Pregnancy, Infection, and Nutrition (PIN) cohort which found lower scores at 2-3 years old on the Mullen Scales of Early Learning Cognitive Composite Scores, specifically the Fine Motor and Expressive Language Scales, with higher prenatal concentrations of isopropyl-phenyl phenyl phosphate (ip-PPP), but no statistically significant associations between DPHP and BDCIPP and child cognition. 28 A study by Castorina et al. conducted among 310 predominantly Hispanic and low-income participants in the CHAMACOS study, found that higher prenatal DPHP concentrations were associated with decreased Full-Scale IQ and Working Memory among children at seven years old. 45 Another study by Liu et al. found significant associations between higher prenatal BDCIPP and lower Mental Development Index at 2 years old, but no significant associations with DPHP. 40 135 Another study by Percy et al. among the HOME study participants found a modest increase in child full- scale IQ at 8 years old with prenatal BCEP concentrations, but no other associations between prenatal BDCIPP, DPHP, and DNBP and child cognition. 51 Evidence of effect modification by child sex also varied by study, with evidence of sex-specific effects in the PIN and Lui et al. studies, but no evidence of sex- specific effects in the CHAMACOS and HOME studies. A variety of factors may have contributed to the discrepancy in findings across studies. For one, concentrations of OPE metabolites varied across cohorts, with higher median DPHP among the HOME, CHAMACOS, and PIN cohorts, but lower median concentrations of DPHP among the Liu et al. study when compared to our study (0.78 ng/mL). Additionally, repeated exposure measurements during multiple windows in pregnancy were only conducted for the Liu et al. study and the HOME study, while the PIN, CHAMACOS, and our study only measured OPE exposure at a single timepoint in pregnancy, with single exposure measures likely resulting in higher exposure misclassification. Furthermore, there was variability in the OPE metabolites measured across studies, with no other study prior to ours measuring prenatal exposures to BMPP, DPRP, BEHP, and BCIPP. Additional differences in gestational age at sample collection, years of sample collection, tools used to assess neurodevelopmental outcomes, and children’s ages at outcome assessment could have contributed to the observed heterogeneity across studies. Current toxicological and epidemiological evidence supports the potential for prenatal OPEs to adversely impact motor development through multiple mechanisms, including perturbations of glutamate and GABA neurotransmitters, 52-58 inflammation, 55,59 glia activation, 60,61 oxidative stress, 55,57,62 and decreased neuronal growth, cell differentiation, and network activity. 52,63-65 Experimental evidence additionally suggests that OPEs may impact cellular apoptosis and may be cytotoxic to neuronal cells. 15,66,67 Additional hypothesized mechanisms include indirect impacts on motor development via endocrine- disrupting pathways, which play a critical role in supporting healthy cognitive and motor development, with prior epidemiological evidence finding associations between OPE exposures and altered levels of thyroid stimulating hormone (TSH) 68 and disruption of other thyroid hormones, 69 along with disruption of sex-steroid hormones and sex-steroid binding globulins. 70-72 136 This study has several important strengths. For one, the prospective design of this study provided us with the opportunity to collect urine samples during pregnancy, a potentially sensitive period of development, prior to assessing our outcome of interest. Additionally, since maternal urinary measures are considered reliable indicators of potential fetal OPE exposures, an additional strength of this study was the use of prenatal urinary metabolites as a measure of in utero exposure to OPEs. 20 We also evaluated various previously understudied OPE metabolites, including DNBP+DIBP, BCIPP, BCEP, BBOEP, DRPR, BMPP, and BEHP, which advances risk assessment opportunities and subsequent targeted interventions. Furthermore, the population evaluated in this study was largely comprised of pregnant individuals born in Latin America, who are disproportionally burdened by environmental exposures and historically underrepresented in U.S. biomedical and population health research, providing us with the opportunity to inform environmental justice solutions. Our study also has some limitations. Since single spot urine samples collected during the third trimester were used to assess OPE exposures throughout pregnancy, there may have been some exposure misclassification. Previous studies indicate moderate to good reproducibility for DPHP and BDCIPP levels throughout pregnancy, but very little is known about the reproducibility of many of the understudied OPE metabolites included in this study. 26,73 Additionally, although we adjusted for many key covariates, residual confounding could still be present, especially for postnatal OPE exposures, which could impact early motor development outcomes. CONCLUSION Our results suggest adverse effects of prenatal OPE exposures on fine motor development and sex- specific impacts of OPE metabolites on gross and fine motor development, with more adverse impacts among female infants. Future research examining multiple windows of prenatal and postnatal susceptibility to OPE exposures is warranted. 137 Figure 1: Consort Diagram of Included Mother-Infant Dyads. Available ASQ-3 data from least one timepoint (n=332) Final sample with complete data on OPEs and ASQ-3 (n=329) Missing fine/gross motor scale ASQ-3 at 6, 9, 12, and 18 months (n=81) Withdrawn from study (n=13) Missing data on key covariates (n=3) Participants with prenatal OPE metabolite concentrations (n=426) 138 Figure 2: Spearman Correlations of Organophosphate Ester Metabolite Concentrations in Third Trimester Maternal Urine. Note: DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 139 Figure 3: Distribution of Ages and Stages (ASQ-3) Gross and Fine Motor Subscale Scores by Study Timepoint (N=329) From top right to left: Median (IQR) gross motor ASQ-3 scores= 45 (15), 50 (30), 50 (20), 60 (10). From bottom right to left: Median (IQR) fine motor ASQ-3 scores= 55 (20), 55 (10), 55 (10), 55 (10). Note: Scores below the red dashed line are 2 standard deviations away from the average scores of a normative US sample, with lower scores suggestive of concern. Scores between the blue line and dashed red line are 1 standard deviations below the average scores of a normative US sample. 140 Table 1: Participant Characteristics (N=329) Mean (SD)/Freq(%) Maternal Characteristics Age (years) 29.2 (6.1) Education Less than or equal to 12 th grade More than 12 th grade 186 (56.5%) 143 (43.5%) Income Don’t Know Less than $30,000 $30,000+ 96 (29.2%) 153 (46.5%) 80 (24.3%) NIH Race Categories White, non-Hispanic Black, non-Hispanic Hispanic Multiracial/Other non-Hispanic 24 (7.3%) 33 (10.0%) 267 (81.2%) 5 (1.5%) Smoking During Pregnancy No Yes 322 (97.9%) 7 (2.1%) Pre-pregnancy BMI (kg/m 2 ) 29.9 (6.6) Infant Characteristics Sex Female Male 170 (51.7%) 159 (48.3%) Birth Order First-Born Second Born or More Missing 110 (33.4%) 207 (62.9%) 12 (3.7%) Gestational Age at Birth (weeks) 39.1 (1.5) Corrected Age (weeks) at Administration a 6-month questionnaire 9-month questionnaire 12-month questionnaire 18-month questionnaire 28.0 (3.7) 40.8 (1.8) 51.0 (2.5) 80.2 (1.5) a Infant age at questionnaire administration was corrected for preterm birth (<37 weeks). 141 Table 2. Distribution of Specific Gravity Adjusted OPEs Concentrations (ng/mL) in Maternal Urine (N=329) Percentiles Distributions Metabolite 25th 50th 75th Min-Max Geometric Mean Detect Frequency LOD (ng/mL) DPHP 0.48 0.78 1.47 0.14-25.59 0.89 99.70% 0.0281 DNBP+DIBP 0.12 0.17 0.26 ND-3.01 0.19 97.87% 0.0441 BDCIPP 0.64 1.32 2.29 ND-26.19 1.06 93.92% 0.0174 BCEP 0.02 0.56 1.92 ND-168.00 0.34 69.91% 0.0200 BBOEP ND 0.04 0.07 ND-1.67 0.04 65.35% 0.0199 BCIPP ND 0.18 0.73 ND-19.90 0.13 54.71% 0.0204 BMPP ND 0.01 0.03 ND-0.69 0.02 38.91% 0.0115 BEHP ND ND 0.03 ND-4.42 0.03 25.23% 0.0170 DPRP ND ND 0.05 ND-2.85 0.03 26.14% 0.0278 ∑OPEs (nmol/L) 10.48 17.75 31.81 2.48-765.32 18.91 Note: OPE, Organophosphate Ester; LOD, Below Limit of Detection; ND, Not Detected; Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 142 Table 3: Longitudinal Associations Between Prenatal OPEs and ASQ-3 Motor Development Across Infancy (N=329) ASQ-3 Motor Development Groups (“High Risk of Delay” vs “Typically developing”) Reversed Scored ASQ-3 Counts‡ Metabolites Gross Motor Outcome Models OR (95% CI) Fine Motor Outcome Models OR (95% CI) Gross Motor Outcome Models IRR (95% CI) Fine Motor Outcome Models IRR (95% CI) DPHP† 1.05 (0.84, 1.31) 1.26 (1.02, 1.57)* 1.04 (0.96, 1.12) 1.11 (1.01, 1.22)* DNBP+DIBP† 1.30 (0.96, 1.74) 1.02 (0.75, 1.37) 1.08 (0.97, 1.19) 1.02 (0.90, 1.15) BDCIPP† 1.03 (0.90, 1.18) 1.00 (0.87, 1.14) 1.02 (0.98, 1.07) 0.99 (0.94, 1.04) BCEP† 0.99 (0.91, 1.08) b 1.00 (0.92, 1.09) 1.00 (0.97, 1.03) 1.02 (0.98, 1.05) BBOEP† 1.07 (0.86, 1.32) 1.11 (0.90, 1.36) 1.00 (0.93, 1.08) 1.04 (0.96, 1.13) BCIPP Not Detected Detected REF 0.86 (0.49, 1.52) REF 1.34 (0.76, 2.38) a REF 0.93 (0.76, 1.13) REF 1.09 (0.86, 1.37) BMPP Not Detected Detected REF 0.60 (0.33, 1.11) REF 1.02 (0.57, 1.81) REF 0.77 (0.63, 0.94)* REF 1.12 (0.88, 1.41) BEHP Not Detected Detected REF 0.92 (0.46, 1.83) REF 0.85 (0.42, 1.72) REF 0.96 (0.77, 1.21) REF 0.99 (0.75, 1.32) DPRP Not Detected Detected REF 0.87 (0.45, 1.69) a REF 1.31 (0.69, 2.50) REF 0.94 (0.75, 1.17) a REF 1.23 (0.95, 1.59) ∑OPEs† 1.06 (0.85, 1.32) 1.02 (0.82, 1.27) 1.04 (0.96, 1.12) 1.04 (0.95, 1.14) Models adjusted for recruitment site, maternal age, GA at sample collection, season of sample collection, infant corrected age at questionnaire administration, infant birth order, pre-pregnancy BMI, race/ethnicity, household income, education, infant sex. Note: OPE, Organophosphate Ester; ASQ-3, Ages and Stages Questionnaire, Third Edition; CI, Confidence Interval; REF, Reference; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3- dichloro-2-propyl) phosphate; BCEP, Bis (2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2- propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. †Beta estimate back transformed to a doubling in exposure prior to exponentiation. a Sex interaction term p-value <0.05. b Time interaction term p-value <0.05. ‡ Since counts are reverse scored, a positive IRR is indicative of poorer performance on the respective motor scale. 143 Table 4: Longitudinal Associations Between Prenatal OPEs and ASQ-3 Motor Development Groups (“High Risk of Delay” vs “Typically developing”) Across Infancy, by Infant Sex Gross Motor Outcome Models Fine Motor Outcome Models Metabolites Female Only (N=170) OR (95% CI) Male Only (N=159) OR (95% CI) Female Only (N=170) OR (95% CI) Male Only (N=159) OR (95% CI) DPHP† 1.05 (0.81, 1.34) 1.09 (0.65, 1.83) 1.15 (0.88, 1.49) 1.46 (1.01, 2.11)* DNBP+DIBP† 1.00 (0.70, 1.45) 2.06 (0.82, 5.21) 0.97 (0.65, 1.46) 0.97 (0.59, 1.60) BDCIPP† 1.04 (0.90, 1.22) 1.09 (0.81, 1.47) 1.02 (0.86, 1.20) 0.95 (0.76, 1.20) BCEP† 0.99 (0.89, 1.10) 0.95 (0.79, 1.14) 1.02 (0.91, 1.14) 0.98 (0.86, 1.12) BBOEP† 1.06 (0.82, 1.39) 1.06 (0.67, 1.70) 0.97 (0.72, 1.30) 1.36 (0.99, 1.87) BCIPP Not Detected Detected REF 1.07 (0.91, 1.26) REF 0.97 (0.72, 1.31) REF 2.72 (1.27, 5.85)* REF 0.76 (0.31, 1.90) BMPP Not Detected Detected REF 0.63 (0.30, 1.33) REF 0.44 (0.11, 1.83) REF 0.75 (0.35, 1.64) REF 1.55 (0.64, 3.78) BEHP Not Detected Detected REF 0.84 (0.37, 1.94) REF 1.41 (0.33, 6.07) REF 1.23 (0.50, 3.01) REF 0.71 (0.23, 2.18) DPRP Not Detected Detected REF 1.48 (0.71, 3.09) REF 0.27 (0.06, 1.29) REF 1.09 (0.47, 2.51) REF 1.25 (0.44, 3.51) ∑OPEs† 1.01 (0.77, 1.31) 1.15 (0.73, 1.80) 1.03 (0.78, 1.38) 1.06 (0.75, 1.51) Models adjusted for recruitment site, maternal age, GA at sample collection, season of sample collection, infant corrected age at questionnaire administration, infant birth order, pre-pregnancy BMI, race/ethnicity, household income, education. Note: OPE, Organophosphate Ester; ASQ-3, Ages and Stages Questionnaire, Third Edition; CI, Confidence Interval; REF, Reference; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3- dichloro-2-propyl) phosphate; BCEP, Bis (2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2- propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. †Beta estimate back transformed to a doubling in exposure prior to exponentiation. 144 Table 5: Longitudinal Associations Between Prenatal OPEs and ASQ-3 Motor Development Counts, by Infant Sex Gross Motor Outcome Models Fine Motor Outcome Models Metabolites Female Only (N=170) OR (95% CI) Male Only (N=159) OR (95% CI) Female Only (N=170) OR (95% CI) Male Only (N=159) IRR (95% CI) DPHP† 1.03 (0.92, 1.14) 1.04, 0.93, 1.17) 1.08 (0.95, 1.23) 1.10 (0.95, 1.27) DNBP+DIBP† 1.00 (0.86, 1.16) 1.14 (0.99, 1.31) 1.00 (0.84, 1.19) 0.99 (0.83, 1.20) BDCIPP† 1.00 (0.94, 1.07) 1.06 (0.99, 1.13) 0.96 (0.90, 1.03) 1.02 (0.94, 1.11) BCEP† 1.01 (0.97, 1.06) 0.98 (0.94, 1.02) 1.04 (0.98, 1.09) 1.00 (0.95, 1.05) BBOEP† 0.99 (0.88, 1.10) 0.99 (0.90, 1.09) 1.02 (0.89, 1.17) 1.05 (0.93, 1.19) BCIPP Not Detected Detected REF 1.04 (0.79, 1.37) REF 0.86 (0.67, 1.12) REF 1.27 (0.92, 1.76) REF 1.01 (0.71, 1.42) BMPP Not Detected Detected REF 0.74 (0.55, 0.99)* REF 0.78 (0.60, 1.02) REF 1.01 (0.72, 1.41) REF 1.25 (0.89, 1.74) BEHP Not Detected Detected REF 0.88 (0.63, 1.24) REF 1.14 (0.84, 1.55) REF 1.06 (0.71, 1.57) 0.99 (0.66, 1.48) DPRP Not Detected Detected REF 1.29 (0.94, 1.77) REF 0.63 (0.47, 0.86)* REF 1.23 (0.86, 1.78) 1.11 (0.76, 1.60) ∑OPEs† 1.01 (0.90, 1.13) 1.07 (0.97, 1.18) 1.03 (0.90, 1.18) 1.06 (0.92, 1.21) Models adjusted for recruitment site, maternal age, GA at sample collection, season of sample collection, infant corrected age at questionnaire administration, infant birth order, pre-pregnancy BMI, race/ethnicity, household income, education. Note: OPE, Organophosphate Ester; ASQ-3, Ages and Stages Questionnaire, Third Edition; CI, Confidence Interval; REF, Reference; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3- dichloro-2-propyl) phosphate; BCEP, Bis (2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2- propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. †Beta estimate back transformed to a doubling in exposure prior to exponentiation. a Since counts are reverse scored, a positive IRR is indicative of poorer performance on the respective motor scale. 145 SUPPLEMENTAL MATERIALS Supplemental Figure 1: Directed Acyclic Graph (DAG) of Prenatal OPE Metabolites and Infant Motor Development Directed Acyclic Graph (DAG) used to identify potential confounders and precision variables. The DAG was created using DAGitty. Green ovals represent exposures or predictors of the exposure, pink ovals represent potential confounders, and blue ovals represent the outcome or predictors of outcome. Minimally sufficient set: Maternal age, parity, pre-pregnancy BMI, prenatal smoking, race/ethnicity, socioeconomic status (SES) 146 Supplemental Figure 2: Associations Between Prenatal OPE Metabolite Concentrations (ng/mL) and Gross and Fine Motor ASQ-3 Groups Models adjusted for recruitment site, maternal age, GA at sample collection, season of sample collection, infant corrected age at questionnaire administration, infant birth order, pre-pregnancy BMI, race/ethnicity, household income, education, infant sex. Note: OPE, Organophosphate Ester; ASQ-3, Ages and Stages Questionnaire, Third Edition; CI, Confidence Interval; REF, Reference; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. *p<0.05 147 Supplemental Figure 3: Associations Between Prenatal OPE Metabolite Concentrations (ng/mL) and Gross and Fine Motor ASQ-3 Counts Models adjusted for recruitment site, maternal age, GA at sample collection, season of sample collection, infant corrected age at questionnaire administration, infant birth order, pre-pregnancy BMI, race/ethnicity, household income, education, infant sex. Note: OPE, Organophosphate Ester; ASQ-3, Ages and Stages Questionnaire, Third Edition; CI, Confidence Interval; REF, Reference; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis (2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. *p<0.05 148 Supplemental Table 1: Comparison of Participant Characteristics in the Full Analytical Sample vs those Excluded from the Sample Due to Missing Information on Key Variables Full Analytical Sample (N=329) Mean (SD)/Freq(%) Excluded From Sample (N= 97) Mean (SD)/Freq(%) Maternal Characteristics Age (years) 29.2 (6.1) 27.7 (5.8) Education Less than 12 th grade More than 12 th grade Missing 186 (56.5%) 143 (43.5%) ─ 54 (55.7%) 39 (40.2%) 4 (4.1%) Income Don’t Know Less than $30,000 $30,000+ Missing 96 (29.2%) 153 (46.5%) 80 (24.3%) ─ 26 (26.8%) 47 (48.5%) 20 (20.6%) 4 (4.1%) NIH Race Categories White, non-Hispanic Black, non-Hispanic Hispanic Multiracial/Other non-Hispanic Missing 24 (7.3%) 33 (10.0%) 267 (81.2%) 5 (1.5%) ─ 5 (5.2%) 16 (16.5%) 62 (63.9%) 10 (10.3%) 4 (4.1%) Smoking During Pregnancy No Yes 322 (97.9%) 7 (2.1%) 96 (99.0%) 1 (1.0%) Pre-pregnancy BMI (kg/m 2 ) 29.9 (6.6) 27.5 (6.8) Infant Characteristics Sex Female Male 170 (51.7%) 159 (48.3%) 48 (49.5%) 49 (50.5%) Infant Birth Order First-Born Second Born or More Missing 110 (33.4%) 207 (62.9%) 12 (3.7%) 37 (38.1%) 54 (55.7%) 6 (6.2%) Gestational Age at Birth (weeks) 39.1 (1.5) 39.0 (1.5) 149 Supplemental Table 2: Cross-Sectional Associations Between Prenatal OPEs and ASQ-3 Motor Development Groups (“Typically Developing” versus “High Risk of Delay” versus) Across Infancy 6-Months (N=311) 12-Months (N=185) Metabolites Gross Motor Outcome Models OR (95% CI) Fine Motor Outcome Models OR (95% CI) Gross Motor Outcome Models OR (95% CI) Fine Motor Outcome Models OR (95% CI) DPHP† 1.05 (0.83, 1.33) 1.30 (1.01, 1.67)* 1.20 (0.89, 1.61) 1.12 (0.76, 1.64) DNBP+DIBP† 1.20 (0.89, 1.62) 1.01 (0.71, 1.44) 1.27 (0.88, 1.83) 1.06 (0.64, 1.75) BDCIPP† 1.14 (0.97, 1.34) 1.04 (0.88, 1.23) 0.95 (0.81, 1.11) 0.90 (0.74, 1.10) BCEP† 0.98 (0.90, 1.07) 1.00 (0.90, 1.11) 1.04 (0.93, 1.17) 1.04 (0.91, 1.20) BBOEP† 1.20 (0.96, 1.51) 0.90 (0.69, 1.17) 0.80 (0.60, 1.06) 1.13 (0.81, 1.57) BCIPP Not Detected Detected REF 1.33 (0.73, 2.43) REF 1.52 (0.76, 3.05) REF 0.70 (0.35, 1.42) REF 1.67 (0.66, 4.22) BMPP Not Detected Detected REF 0.82 (0.44, 1.53) REF 1.04 (0.52, 2.09) REF 0.58 (0.27, 1.21) REF 0.77 (0.30, 1.97) BEHP Not Detected Detected REF 0.93 (0.46, 1.88) REF 0.96 (0.42, 2.15) REF 1.38 (0.57, 3.36) REF 0.56 (0.16, 1.98) DPRP Not Detected Detected REF 1.00 (0.51, 1.95) 1.43 (0.68, 3.01) REF 0.86 (0.37, 2.04) REF 0.74 (0.24, 2.27) ∑OPEs† 1.07 (0.85, 1.34) 0.99 (0.76, 1.29) 1.15 (0.86, 1.53) 1.19 (0.83, 1.70) Models adjusted for recruitment site, maternal age, GA at sample collection, season of sample collection, infant corrected age at questionnaire administration, infant birth order, pre-pregnancy BMI, race/ethnicity, household income, education, infant sex. Note: OPE, Organophosphate Ester; ASQ-3, Ages and Stages Questionnaire, Third Edition; CI, Confidence Interval; REF, Reference; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2- propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. †Beta estimate back transformed to a doubling in exposure. 150 Supplemental Table 3: Cross-Sectional Associations Between Prenatal OPEs and ASQ-3 Motor Counts Across Infancy 6-Months (N=311) 12-Months (N=185) Metabolites Gross Motor Outcome Models IRR (95% CI) Fine Motor Outcome Models IRR (95% CI) Gross Motor Outcome Models IRR (95% CI) Fine Motor Outcome Models IRR (95% CI) DPHP† 1.05 (0.97, 1.13) 1.14 (1.01, 1.29)* 1.13 (0.99, 1.29) 1.05 (0.89, 1.23) DNBP+DIBP† 1.04 (0.94, 1.15) 1.01 (0.85, 1.20) 1.18 (0.99, 1.40) 0.99 (0.82, 1.21) BDCIPP† 1.04 (1.00, 1.09) 0.99 (0.92, 1.07) 1.01 (0.94, 1.09) 0.95 (0.87, 1.03) BCEP† 1.00 (0.97, 1.03) 1.01 (0.96, 1.06) 1.03 (0.98, 1.08) 1.03 (0.97, 1.10) BBOEP† 1.01 (0.94, 1.09) 1.01 (0.89, 1.15) 0.97 (0.86, 1.10) 1.01 (0.88, 1.16) BCIPP Not Detected Detected REF 1.11 (0.92, 1.34) REF 1.06 (0.77, 1.48) REF 0.81 (0.58, 1.13) REF 1.35 (0.92, 1.99) BMPP Not Detected Detected REF 0.84 (0.69, 1.02) REF 1.38 (1.00, 1.92) REF 0.80 (0.58, 1.12) REF 0.91 (0.62, 1.35) BEHP Not Detected Detected REF 0.93 (0.75, 1.15) REF 0.92 (0.63, 1.34) REF 1.07 (0.71, 1.62) REF 0.88 (0.54, 1.45) DPRP Not Detected Detected REF 0.98 (0.80, 1.21) REF 1.21 (0.84, 1.75) 0.93 (0.62, 1.37) REF 0.93 (0.59, 1.45) ∑OPEs† 1.04 (0.96, 1.12) 1.07 (0.94, 1.22) 1.10 (0.97, 1.25) 1.05 (0.90, 1.22) Models adjusted for recruitment site, maternal age, GA at sample collection, season of sample collection, infant corrected age at questionnaire administration, infant birth order, pre-pregnancy BMI, race/ethnicity, household income, education, infant sex. Note: OPE, Organophosphate Ester; CBCL, Child Behavior Check List; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2- ethylhexyl) phosphate; DPRP, Dipropyl phosphate. †Beta estimate back transformed to a doubling in exposure. 151 Supplemental Table 4: Longitudinal Associations Between Prenatal OPEs and ASQ-3 Motor Development Across Infancy Excluding Participants that Smoked In-Utero (N=329) ASQ-3 Motor Development Groups (“Typical” vs “High Risk of Delay”) Reversed Scored ASQ-3 Counts Metabolites Gross Motor Outcome Models OR (95% CI) Fine Motor Outcome Models OR (95% CI) Gross Motor Outcome Models IRR (95% CI) Fine Motor Outcome Models IRR (95% CI) DPHP† 1.05 (0.84, 1.32) 1.28 (1.03, 1.60)* 1.03 (0.95, 1.12) 1.11 (1.01, 1.22)* DNBP+DIBP† 1.29 (0.95, 1.73) 1.04 (0.77, 1.41) 1.07 (0.97, 1.19) 1.03 (0.91, 1.17) BDCIPP† 1.02 (0.89, 1.17) 1.00 (0.87, 1.15) 1.02 (0.97, 1.07) 1.00 (0.94, 1.05) BCEP† 1.00 (0.92, 1.09) 1.00 (0.92, 1.09) 1.00 (0.97, 1.03) 1.01 (0.98, 1.05) BBOEP† 1.07 (0.86, 1.32) 1.09 (0.88, 1.34) 0.99 (0.92, 1.07) 1.03 (0.94, 1.12) BCIPP Not Detected Detected REF 0.83 (0.47, 1.46) REF 1.36 (0.76, 2.43) REF 0.91 (0.74, 1.10) REF 1.09 (0.86, 1.38) BMPP Not Detected Detected REF 0.63 (0.34, 1.16) REF 0.97 (0.54, 1.74) REF 0.78 (0.64, 0.96)* REF 1.08 (0.85, 1.37) BEHP Not Detected Detected REF 0.98 (0.49, 1.98) REF 0.72 (0.35, 1.52) REF 0.97 (0.77, 1.23) REF 0.92 (0.69, 1.23) DPRP Not Detected Detected REF 0.90 (0.46, 1.75) REF 1.29 (0.67, 2.47) REF 0.95 (0.76, 1.19) REF 1.21, 0.93, 1.56) ∑OPEs 1.05 (0.84, 1.32) 1.00 (0.80, 1.26) 1.03 (0.95, 1.11) 1.03 (0.94, 1.13) Models adjusted for recruitment site, maternal age, GA at sample collection, season of sample collection, infant corrected age at questionnaire administration, infant birth order, pre-pregnancy BMI, race/ethnicity, household income, education, infant sex. Note: OPE, Organophosphate Ester; ASQ-3, Ages and Stages Questionnaire, Third Edition; CI, Confidence Interval; REF, Reference; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3- dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2- propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. †Beta estimate back transformed to a doubling in exposure. 152 Supplemental Table 5: Longitudinal Associations Between Prenatal OPEs and ASQ-3 Motor Development Across Infancy Among Full Term Births Only (≥37 weeks GA at Birth) (N=301) ASQ-3 Motor Development Groups (“Typical” vs “High Risk of Delay”) Reversed Scored ASQ-3 Counts Metabolites Gross Motor Outcome Models OR (95% CI) Fine Motor Outcome Models OR (95% CI) Gross Motor Outcome Models IRR (95% CI) Fine Motor Outcome Models IRR (95% CI) DPHP† 1.10 (0.87, 1.39) 1.37 (1.08, 1.74)* 1.05 (0.97, 1.14) 1.13 (1.02, 1.25)* DNBP+DIBP† 1.29 (0.95, 1.76) 0.95 (0.69, 1.33) 1.07 (0.95, 1.19) 0.97 (0.85, 1.11) BDCIPP† 1.03 (0.90, 1.17) 0.99 (0.86, 1.13) 1.02 (0.98, 1.07) 0.99 (0.93, 1.04) BCEP† 0.97 (0.88, 1.06) 0.99 (0.90, 1.08) 0.99 (0.96, 1.02) 1.01 (0.97, 1.05) BBOEP† 1.15 (0.91, 1.44) 1.04 (0.82, 1.30) 1.01 (0.94, 1.10) 1.04 (0.94, 1.14) BCIPP Not Detected Detected REF 0.87 (0.48, 1.56) REF 1.37 (0.74, 2.53) REF 0.97 (0.79, 1.20) REF 1.08 (0.85, 1.38) BMPP Not Detected Detected REF 0.70 (0.37, 1.31) REF 1.20 (0.65, 2.21) REF 0.83 (0.67, 1.02) REF 1.20 (0.94, 1.54) BEHP Not Detected Detected REF 0.98 (0.49, 1.97) REF 0.86 (0.41, 1.79) REF 0.98 (0.77, 1.25) REF 1.02 (0.76, 1.36) DPRP Not Detected Detected REF 0.80 (0.39, 1.61) REF 1.55 (0.77, 3.10) REF 0.97 (0.76, 1.23) REF 1.29 (0.97, 1.70) ∑OPEs 1.00 (0.79, 1.27) 0.98 (0.77, 1.25) 1.01 (0.93, 1.10) 1.03 ( 0.93, 1.14) Models adjusted for recruitment site, maternal age, GA at sample collection, season of sample collection, infant corrected age at questionnaire administration, infant birth order, pre-pregnancy BMI, race/ethnicity, household income, education, infant sex. Note: OPE, Organophosphate Ester; ASQ-3, Ages and Stages Questionnaire, Third Edition; CI, Confidence Interval; REF, Reference; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3- dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2- propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. †Beta estimate back transformed to a doubling in exposure. 153 Supplemental Table 6: Longitudinal Associations Between Prenatal OPEs and ASQ-3 Motor Development Across Infancy, When Rounding Non-Integers Up vs Down (N=329) Rounding Up Rounding Down Metabolites Gross Motor Outcome Models IRR (95% CI) Fine Motor Outcome Models IRR (95% CI) Gross Motor Outcome Models IRR (95% CI) Fine Motor Outcome Models IRR (95% CI) DPHP† 1.04 (0.96, 1.12) 1.11 (1.01, 1.22)* 1.04 (0.96, 1.12) 1.11 (1.01, 1.21)* DNBP+DIBP† 1.07 (0.97, 1.19) 1.02 (0.90, 1.15) 1.08 (0.97, 1.19) 1.02 (0.91, 1.15) BDCIPP† 1.02 (0.98, 1.07) 0.99 (0.94, 1.04) 1.02 (0.98, 1.07) 0.99 (0.94, 1.04) BCEP† 1.00 (0.97, 1.03) 1.02 (0.98, 1.05) 1.00 (0.97, 1.03) 1.02 (0.98, 1.05) BBOEP† 1.00 (0.93, 1.08) 1.04 (0.96, 1.13) 1.00 (0.93, 1.08) 1.04 (0.96, 1.14) BCIPP Not Detected Detected REF 0.93 (0.76, 1.13) REF 1.09 (0.87, 1.38) REF 0.93 (0.76, 1.13) REF 1.08 (0.86, 1.36) BMPP Not Detected Detected REF 0.77 (0.63, 0.94)* REF 1.12 (0.89, 1.42) REF 0.77 (0.63, 0.94)* REF 1.11 (0.88, 1.41) BEHP Not Detected Detected REF 0.96 (0.76, 1.21) REF 1.00 (0.75, 1.32) REF 0.96 (0.77, 1.21) REF 1.00 (0.75, 1.32) DPRP Not Detected Detected REF 0.94 (0.75, 1.17) REF 1.23 (0.95, 1.59) REF 0.94 (0.75, 1.17) REF 1.22 (0.95, 1.58) ∑OPEs† 1.04 (0.96, 1.12) 1.04 (0.95, 1.14) 1.04 (0.96, 1.12) 1.04 (0.95, 1.14) Models adjusted for recruitment site, maternal age, GA at sample collection, season of sample collection, infant corrected age at questionnaire administration, infant birth order, pre-pregnancy BMI, race/ethnicity, household income, education, infant sex. Note: OPE, Organophosphate Ester; ASQ-3, Ages and Stages Questionnaire, Third Edition; CI, Confidence Interval; REF, Reference; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3- dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2- propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate. †Beta estimate back transformed to a doubling in exposure. 154 REFERENCES 1. Montgomery C, Y FK, Strand Brodd K, et al. Structured Observation of Motor Performance in Infants: Level and quality associated with later motor development. Acta Paediatr 2021; 110: 307-313. 2020/06/01. DOI: 10.1111/apa.15377. 2. Camargos ACR, Mendonça VA, Andrade CAd, et al. Overweight and obese infants present lower cognitive and motor development scores than normal-weight peers. Research in Developmental Disabilities 2016; 59: 410-416. DOI: https://doi.org/10.1016/j.ridd.2016.10.001. 3. Shoaibi A, Neelon B, Østbye T, et al. Longitudinal associations of gross motor development, motor milestone achievement and weight-for-length z score in a racially diverse cohort of US infants. BMJ Open 2019; 9: e024440. DOI: 10.1136/bmjopen-2018-024440. 4. Hewitt L, Kerr E, Stanley RM, et al. Tummy Time and Infant Health Outcomes: A Systematic Review. Pediatrics 2020; 145 2020/05/07. DOI: 10.1542/peds.2019-2168. 5. Stiles J and Jernigan TL. The basics of brain development. Neuropsychol Rev 2010; 20: 327-348. 2010/11/03. DOI: 10.1007/s11065-010-9148-4. 6. Lipkin PH. Chapter 66 - MOTOR DEVELOPMENT AND DYSFUNCTION. In: Carey WB, Crocker AC, Coleman WL, et al. (eds) Developmental-Behavioral Pediatrics (Fourth Edition). Philadelphia: W.B. Saunders, 2009, pp.643-652. 7. Monk C and Hane A. Fetal and Infant Brain–Behavior Development: Milestones & Environmental Influences. 2014. 8. Grandjean P and Landrigan PJ. Neurobehavioural effects of developmental toxicity. The Lancet Neurology 2014; 13: 330-338. DOI: 10.1016/S1474-4422(13)70278-3. 9. Miguel PM, Pereira LO, Silveira PP, et al. Early environmental influences on the development of children's brain structure and function. Dev Med Child Neurol 2019; 61: 1127-1133. 2019/02/12. DOI: 10.1111/dmcn.14182. 10. Rock KD and Patisaul HB. Environmental Mechanisms of Neurodevelopmental Toxicity. Curr Environ Health Rep 2018; 5: 145-157. DOI: 10.1007/s40572-018-0185-0. 11. Thomason ME, Hect J, Waller R, et al. Prenatal neural origins of infant motor development: Associations between fetal brain and infant motor development. Dev Psychopathol 2018; 30: 763-772. 2018/08/03. DOI: 10.1017/s095457941800072x. 12. Brian A, Pennell A, Taunton S, et al. Motor Competence Levels and Developmental Delay in Early Childhood: A Multicenter Cross-Sectional Study Conducted in the USA. Sports Med 2019; 49: 1609-1618. 2019/07/14. DOI: 10.1007/s40279-019-01150-5. 13. Handal AJ, Lozoff B, Breilh J, et al. Effect of community of residence on neurobehavioral development in infants and young children in a flower-growing region of Ecuador. Environ Health Perspect 2007; 115: 128-133. 2007/03/21. DOI: 10.1289/ehp.9261. 14. Guo J, Zhang J, Wu C, et al. Associations of prenatal and childhood chlorpyrifos exposure with Neurodevelopment of 3-year-old children. Environ Pollut 2019; 251: 538-546. 2019/05/21. DOI: 10.1016/j.envpol.2019.05.040. 15. Dishaw LV, Powers CM, Ryde IT, et al. Is the PentaBDE replacement, tris (1,3-dichloro-2- propyl) phosphate (TDCPP), a developmental neurotoxicant? Studies in PC12 cells. Toxicology and applied pharmacology 2011; 256: 281-289. 2011/01/19. DOI: 10.1016/j.taap.2011.01.005. 16. Patisaul HB, Behl M, Birnbaum LS, et al. Beyond Cholinesterase Inhibition: Developmental Neurotoxicity of Organophosphate Ester Flame Retardants and Plasticizers. Environ Health Perspect 2021; 129: 105001. 2021/10/07. DOI: 10.1289/ehp9285. 17. Yang J, Zhao Y, Li M, et al. A Review of a Class of Emerging Contaminants: The Classification, Distribution, Intensity of Consumption, Synthesis Routes, Environmental Effects and Expectation of Pollution Abatement to Organophosphate Flame Retardants (OPFRs). Int J Mol Sci 2019; 20 2019/06/20. DOI: 10.3390/ijms20122874. 155 18. Blum A, Behl M, Birnbaum L, et al. Organophosphate Ester Flame Retardants: Are They a Regrettable Substitution for Polybrominated Diphenyl Ethers? Environ Sci Technol Lett 2019; 6: 638- 649. 2020/06/05. DOI: 10.1021/acs.estlett.9b00582. 19. Bergman A, Rydén A, Law RJ, et al. A novel abbreviation standard for organobromine, organochlorine and organophosphorus flame retardants and some characteristics of the chemicals. Environ Int 2012; 49: 57-82. 2012/09/13. DOI: 10.1016/j.envint.2012.08.003. 20. Hou R, Xu Y and Wang Z. Review of OPFRs in animals and humans: Absorption, bioaccumulation, metabolism, and internal exposure research. Chemosphere 2016; 153: 78-90. DOI: https://doi.org/10.1016/j.chemosphere.2016.03.003. 21. Stapleton HM, Klosterhaus S, Eagle S, et al. Detection of organophosphate flame retardants in furniture foam and U.S. house dust. Environ Sci Technol 2009; 43: 7490-7495. 2009/10/24. DOI: 10.1021/es9014019. 22. Du J, Li H, Xu S, et al. A review of organophosphorus flame retardants (OPFRs): occurrence, bioaccumulation, toxicity, and organism exposure. Environ Sci Pollut Res Int 2019; 26: 22126-22136. 2019/06/28. DOI: 10.1007/s11356-019-05669-y. 23. He R, Li Y, Xiang P, et al. Organophosphorus flame retardants and phthalate esters in indoor dust from different microenvironments: Bioaccessibility and risk assessment. Chemosphere 2016; 150: 528- 535. 2015/11/21. DOI: 10.1016/j.chemosphere.2015.10.087. 24. Pantelaki I and Voutsa D. Organophosphate flame retardants (OPFRs): A review on analytical methods and occurrence in wastewater and aquatic environment. Sci Total Environ 2019; 649: 247-263. 2018/09/03. DOI: 10.1016/j.scitotenv.2018.08.286. 25. Ospina M, Jayatilaka NK, Wong LY, et al. Exposure to organophosphate flame retardant chemicals in the U.S. general population: Data from the 2013-2014 National Health and Nutrition Examination Survey. Environ Int 2018; 110: 32-41. 2017/11/06. DOI: 10.1016/j.envint.2017.10.001. 26. Hoffman K, Daniels JL and Stapleton HM. Urinary metabolites of organophosphate flame retardants and their variability in pregnant women. Environ Int 2014; 63: 169-172. 2013/12/10. DOI: 10.1016/j.envint.2013.11.013. 27. Castorina R, Butt C, Stapleton HM, et al. Flame retardants and their metabolites in the homes and urine of pregnant women residing in California (the CHAMACOS cohort). Chemosphere 2017; 179: 159- 166. 2017/03/22. DOI: 10.1016/j.chemosphere.2017.03.076. 28. Doherty BT, Hoffman K, Keil AP, et al. Prenatal exposure to organophosphate esters and cognitive development in young children in the Pregnancy, Infection, and Nutrition Study. Environ Res 2019; 169: 33-40. 2018/11/09. DOI: 10.1016/j.envres.2018.10.033. 29. Kuiper JR, Stapleton HM, Wills-Karp M, et al. Predictors and reproducibility of urinary organophosphate ester metabolite concentrations during pregnancy and associations with birth outcomes in an urban population. Environ Health 2020; 19: 55. 2020/05/26. DOI: 10.1186/s12940-020-00610-0. 30. Thomas MB, Stapleton HM, Dills RL, et al. Demographic and dietary risk factors in relation to urinary metabolites of organophosphate flame retardants in toddlers. Chemosphere 2017; 185: 918-925. 2017/08/03. DOI: 10.1016/j.chemosphere.2017.07.015. 31. Wang X, Zhu Q, Yan X, et al. A review of organophosphate flame retardants and plasticizers in the environment: Analysis, occurrence and risk assessment. Sci Total Environ 2020; 731: 139071. 2020/05/22. DOI: 10.1016/j.scitotenv.2020.139071. 32. Wang Y, Li W, Martínez-Moral MP, et al. Metabolites of organophosphate esters in urine from the United States: Concentrations, temporal variability, and exposure assessment. Environ Int 2019; 122: 213-221. 2018/11/20. DOI: 10.1016/j.envint.2018.11.007. 33. Cooper E, Covaci A, Van Nuijs A, et al. Analysis of the flame retardant metabolites bis (1, 3- dichloro-2-propyl) phosphate (BDCPP) and diphenyl phosphate (DPP) in urine using liquid chromatography–tandem mass spectrometry. Analytical and bioanalytical chemistry 2011; 401: 2123- 2132. 34. Lynn R, Wong K, Garvie-Gould C, et al. Disposition of the flame retardant, tris (1, 3-dichloro-2- propyl) phosphate, in the rat. Drug metabolism and Disposition 1981; 9: 434-441. 156 35. Nomeir A, Kato S and Matthews H. The metabolism and disposition of tris (1, 3-dichloro-2- propyl) phosphate (Fyrol FR-2) in the rat. Toxicology and applied pharmacology 1981; 57: 401-413. 36. Sasaki K, Suzuki T, Takeda M, et al. Metabolism of phosphoric acid triesters by rat liver homogenate. Bulletin of environmental contamination and toxicology 1984; 33: 281-288. 37. Van den Eede N, Maho W, Erratico C, et al. First insights in the metabolism of phosphate flame retardants and plasticizers using human liver fractions. Toxicology letters 2013; 223: 9-15. 38. Wang G, Du Z, Chen H, et al. Tissue-specific accumulation, depuration, and transformation of triphenyl phosphate (TPHP) in adult zebrafish (Danio rerio). Environmental science & technology 2016; 50: 13555-13564. 39. Wiersielis K, Adams S, Yasrebi A, et al. Maternal exposure to organophosphate flame retardants alters locomotor and anxiety-like behavior in male and female adult offspring. Hormones and behavior 2020; 122: 104759. 40. Liu W, Luo D, Xia W, et al. Prenatal exposure to halogenated, aryl, and alkyl organophosphate esters and child neurodevelopment at two years of age. Journal of Hazardous Materials 2021; 408: 124856. DOI: https://doi.org/10.1016/j.jhazmat.2020.124856. 41. Bastain TM, Chavez T, Habre R, et al. Study Design, Protocol and Profile of the Maternal And Developmental Risks from Environmental and Social Stressors (MADRES) Pregnancy Cohort: a Prospective Cohort Study in Predominantly Low-Income Hispanic Women in Urban Los Angeles. BMC Pregnancy Childbirth 2019; 19: 189-189. DOI: 10.1186/s12884-019-2330-7. 42. Hornung RW and Reed LD. Estimation of Average Concentration in the Presence of Nondetectable Values. Applied Occupational and Environmental Hygiene 1990; 5: 46-51. DOI: 10.1080/1047322X.1990.10389587. 43. Squires J, Twombly,E., Bricker, D., Potter, L. The ASQ-3 Technical Report, https://agesandstages.com/resource/asq-3-technical-appendix/. (2009, 2021). 44. Hoffman K, Lorenzo A, Butt CM, et al. Predictors of urinary flame retardant concentration among pregnant women. Environ Int 2017; 98: 96-101. 2016/10/13. DOI: 10.1016/j.envint.2016.10.007. 45. Castorina R, Bradman A, Stapleton HM, et al. Current-use flame retardants: Maternal exposure and neurodevelopment in children of the CHAMACOS cohort. Chemosphere 2017; 189: 574-580. 2017/10/01. DOI: 10.1016/j.chemosphere.2017.09.037. 46. Textor J, van der Zander B, Gilthorpe MS, et al. Robust causal inference using directed acyclic graphs: the R package 'dagitty'. Int J Epidemiol 2016; 45: 1887-1894. 2017/01/17. DOI: 10.1093/ije/dyw341. 47. Hernandez-Castro I, Eckel SP, Chavez T, et al. Household pesticide exposures and infant gross motor development in the MADRES cohort. Paediatr Perinat Epidemiol 2022; 36: 220-229. 2021/12/30. DOI: 10.1111/ppe.12850. 48. Ghassabian A, Sundaram R, Bell E, et al. Gross motor milestones and subsequent development. Pediatrics 2016; 138. 49. Lingam R, Golding J, Jongmans MJ, et al. The association between developmental coordination disorder and other developmental traits. Pediatrics 2010; 126: e1109-e1118. 50. Lingam R, Jongmans MJ, Ellis M, et al. Mental health difficulties in children with developmental coordination disorder. Pediatrics 2012; 129: e882-e891. 51. Percy Z, Vuong AM, Xu Y, et al. Prenatal exposure to a mixture of organophosphate esters and intelligence among 8-year-old children of the HOME Study. NeuroToxicology 2021; 87: 149-155. DOI: https://doi.org/10.1016/j.neuro.2021.09.005. 52. Hausherr V, van Thriel C, Krug A, et al. Impairment of glutamate signaling in mouse central nervous system neurons in vitro by tri-ortho-cresyl phosphate at noncytotoxic concentrations. Toxicol Sci 2014; 142: 274-284. 2014/09/10. DOI: 10.1093/toxsci/kfu174. 53. Hogberg HT, de Cássia da Silveira ESR, Kleensang A, et al. Organophosphorus flame retardants are developmental neurotoxicants in a rat primary brainsphere in vitro model. Arch Toxicol 2021; 95: 207-228. 2020/10/21. DOI: 10.1007/s00204-020-02903-2. 157 54. Yang W, Zhao F, Fang Y, et al. (1)H-nuclear magnetic resonance metabolomics revealing the intrinsic relationships between neurochemical alterations and neurobehavioral and neuropathological abnormalities in rats exposed to tris(2-chloroethyl)phosphate. Chemosphere 2018; 200: 649-659. 2018/03/09. DOI: 10.1016/j.chemosphere.2018.02.056. 55. Liu X, Zhao X, Wang Y, et al. Triphenyl phosphate permeates the blood brain barrier and induces neurotoxicity in mouse brain. Chemosphere 2020; 252: 126470. 2020/05/24. DOI: 10.1016/j.chemosphere.2020.126470. 56. Gant DB, Eldefrawi ME and Eldefrawi AT. Action of organophosphates on GABAA receptor and voltage-dependent chloride channels. Fundam Appl Toxicol 1987; 9: 698-704. 1987/11/01. DOI: 10.1016/0272-0590(87)90176-x. 57. Wang Q, Lai NL, Wang X, et al. Bioconcentration and transfer of the organophorous flame retardant 1,3-dichloro-2-propyl phosphate causes thyroid endocrine disruption and developmental neurotoxicity in zebrafish larvae. Environ Sci Technol 2015; 49: 5123-5132. 2015/04/01. DOI: 10.1021/acs.est.5b00558. 58. Shi Q, Wang M, Shi F, et al. Developmental neurotoxicity of triphenyl phosphate in zebrafish larvae. Aquat Toxicol 2018; 203: 80-87. 2018/08/11. DOI: 10.1016/j.aquatox.2018.08.001. 59. Zhong X, Wu J, Ke W, et al. Neonatal exposure to organophosphorus flame retardant TDCPP elicits neurotoxicity in mouse hippocampus via microglia-mediated inflammation in vivo and in vitro. Archives of Toxicology 2020; 94: 541-552. DOI: 10.1007/s00204-019-02635-y. 60. Slotkin TA, Skavicus S, Stapleton HM, et al. Brominated and organophosphate flame retardants target different neurodevelopmental stages, characterized with embryonic neural stem cells and neuronotypic PC12 cells. Toxicology 2017; 390: 32-42. DOI: https://doi.org/10.1016/j.tox.2017.08.009. 61. Hogberg HT, de Cássia da Silveira E Sá R, Kleensang A, et al. Organophosphorus flame retardants are developmental neurotoxicants in a rat primary brainsphere in vitro model. Archives of Toxicology 2021; 95: 207-228. DOI: 10.1007/s00204-020-02903-2. 62. Wu S, Ji G, Liu J, et al. TBBPA induces developmental toxicity, oxidative stress, and apoptosis in embryos and zebrafish larvae (Danio rerio). Environmental Toxicology 2016; 31: 1241-1249. https://doi.org/10.1002/tox.22131. DOI: https://doi.org/10.1002/tox.22131. 63. Behl M, Hsieh J-H, Shafer TJ, et al. Use of alternative assays to identify and prioritize organophosphorus flame retardants for potential developmental and neurotoxicity. Neurotoxicology and Teratology 2015; 52: 181-193. DOI: https://doi.org/10.1016/j.ntt.2015.09.003. 64. Ryan KR, Sirenko O, Parham F, et al. Neurite outgrowth in human induced pluripotent stem cell- derived neurons as a high-throughput screen for developmental neurotoxicity or neurotoxicity. NeuroToxicology 2016; 53: 271-281. DOI: https://doi.org/10.1016/j.neuro.2016.02.003. 65. Shafer TJ, Brown JP, Lynch B, et al. Evaluation of Chemical Effects on Network Formation in Cortical Neurons Grown on Microelectrode Arrays. Toxicological Sciences 2019; 169: 436-455. DOI: 10.1093/toxsci/kfz052. 66. Ta N, Li C, Fang Y, et al. Toxicity of TDCPP and TCEP on PC12 cell: changes in CAMKII, GAP43, tubulin and NF-H gene and protein levels. Toxicology letters 2014; 227: 164-171. 67. Pei Y, Peng J, Behl M, et al. Comparative neurotoxicity screening in human iPSC-derived neural stem cells, neurons and astrocytes. Brain research 2016; 1638: 57-73. 68. Yao Y, Li M, Pan L, et al. Exposure to organophosphate ester flame retardants and plasticizers during pregnancy: Thyroid endocrine disruption and mediation role of oxidative stress. Environ Int 2021; 146: 106215. DOI: https://doi.org/10.1016/j.envint.2020.106215. 69. Demeneix Barbara A. Evidence for Prenatal Exposure to Thyroid Disruptors and Adverse Effects on Brain Development. European Thyroid Journal 2019; 8: 283-292. DOI: 10.1159/000504668. 70. Ghassabian A, Henrichs J and Tiemeier H. Impact of mild thyroid hormone deficiency in pregnancy on cognitive function in children: lessons from the Generation R Study. Best practice & research Clinical endocrinology & metabolism 2014; 28: 221-232. 71. Kampouri M, Margetaki K, Koutra K, et al. Maternal mild thyroid dysfunction and offspring cognitive and motor development from infancy to childhood: the Rhea mother-child cohort study in 158 Crete, Greece. J Epidemiol Community Health 2021; 75: 29-35. 2020/09/11. DOI: 10.1136/jech-2019- 213309. 72. Luo K, Liu J, Wang Y, et al. Associations between organophosphate esters and sex hormones among 6-19-year old children and adolescents in NHANES 2013-2014. Environ Int 2020; 136: 105461. 2020/01/14. DOI: 10.1016/j.envint.2020.105461. 73. Romano ME, Hawley NL, Eliot M, et al. Variability and predictors of urinary concentrations of organophosphate flame retardant metabolites among pregnant women in Rhode Island. Environmental Health 2017; 16: 1-11. 159 CHAPTER 5 STUDY 3: Prenatal Exposures to Organophosphate Ester Metabolite Mixtures and Children’s Neurobehavioral Outcomes in the MADRES Pregnancy Cohort ABSTRACT Background: Growing experimental evidence suggests organophosphate esters (OPEs), commonly used as flame retardants and plasticizers, are neurotoxic; however, the epidemiological literature remains scarce. We investigated whether prenatal exposures to OPEs were associated with child neurobehavior in the MADRES cohort. Methods: We measured nine OPE metabolites in 204 maternal urine samples (mean gestational age (GA) at collection: 31.4±1.8 weeks gestation). Neurobehavior problems in child offspring were assessed at 36 months of age using the Child Behavior Checklist's (CBCL) three composite scales [internalizing, externalizing, and total problems]. We examined associations between prenatal metabolite tertiles (>50% detection) and detect/not detect categories (<50% detection) and CBCL composite scales using linear regression and generalized additive models. We also examined OPE mixtures for widely detected OPEs (n=5) using Bayesian Kernel Machine Regression. Results: In single metabolite models, maternal participants with detectable prenatal levels of bis(2-methylphenyl) phosphate (BMPP) were associated with 42% (95% CI: 4%, 96%) higher externalizing and 35% (95% CI: 3%, 78%) higher total problems in children at 36 months when compared to non-detectable prenatal levels of BMPP. Similarly, internalizing scores were 45% (95% CI: -2%, 114%) higher among children with maternal participants who had detectable levels of prenatal BMPP relative to non-detectable levels. Maternal participants in the second tertile of bis(butoxethyl) phosphate (BBOEP) concentrations had children with 43% (95% CI: -1%, 109%) higher externalizing scores when compared to the first tertile of 160 BBOEP exposure; however, children with maternal participants in the third tertile of BBOEP exposure had 13% lower externalizing scores (95% CI: -41%, 29%). A statistically significant interaction term was found between BCIPP and infant sex in internalizing (p=0.02) and total problems (p=0.03) models, with 120% (95% CI: 23%, 295%) and 57% (95% CI: 6%, 134%) higher scores when compared to the first tertile of BCIPP levels among males only. Although associations between the OPE metabolite mixture and each composite CBCL outcome were not statistically significant, we observed a marginal association between dibutyl phosphate and di-isobutyl phosphate (DNBP+DIBP) and higher internalizing scores (0.15; 95% CrI= -0.02, 0.31), holding other metabolites at their median values. In mixture models, BBOEP was ranked as the most important predictor for each outcome, followed by DNBP+DIBP. Conclusion: Our results suggest adverse and sex-specific effects of prenatal exposure to previously understudied OPEs on neurobehavioral symptoms in child offspring at 36 months, providing evidence of potential OPE neurotoxicity. INTRODUCTION Neurobehavioral development is a lifelong, dynamic process which encompasses a host of psychosocial and neurological mechanisms that influence behavior, emotion, and learning. 1,2 Environmental chemical exposures are increasingly recognized as major risk factors for adverse neurobehavioral outcomes, ranging in effects from subclinical deficits in neurobehavioral functioning to increased risks of neurobehavioral disorders. 2-4 The prenatal period is a particularly susceptible window for neurobehavioral development given the rapid cascade of tightly controlled and sequenced neurological processes that occur in utero, resulting in heightened susceptibility to environmental exposures. 2 Even minor, incremental disruptions to prenatal neurological processes from low-level chronic exposures to environmental chemicals have the potential to result in lifelong health effects. 3,5 Flame retardants are anthropogenic chemical additives incorporated into materials to prevent or delay fires and to meet flammability regulations in the United States, particularly in California. 6,7 For many decades, legacy flame retardants, such as polybrominated diphenyl ethers (PBDEs), were the most 161 frequently used. 8,9 However, due to their bioaccumulation in the environment, persistence, and neurotoxicity to children, PBDEs have been phased out of the US market and banned from production in the European Union. 10 As a result, organophosphate esters (OPEs) have dramatically increased in use as replacement flame retardants in recent years. 11-13 However, emerging literature suggests that OPEs may be a regrettable substitution for PBDEs and may also adversely impact neurobehavioral and neurodevelopmental outcomes. 14 OPEs are commonly used as plasticizers and lubricants, contributing to their environmental ubiquity. 7 OPEs are also applied as additives to various consumer, industrial, and electronic products, such as polyurethane foam, textiles, and building materials. 7,15 Due to their physical incorporation within a product matrix and their semivolatile nature, OPEs easily volatize and leach into surrounding environments, commonly settling into dust particles in homes and environmental media such as soil, surface water, sediment, and agricultural products and facilitating human exposure to OPEs. 16-24 As a result, common OPE exposure routes include dermal contact, inhalation, and ingestion of air and dust particles, as well as dietary ingestion of OPE-contaminated food and drinking water. 7,15 OPEs have been found in the placenta and cord blood, suggesting possible in utero transfer to the fetus, and resulting in growing concern, particularly regarding early neurodevelopment, given the structural similarity between OPEs and organophosphate pesticides which have been previously found to be neurotoxic. 25-30 The two most frequently detected OPE metabolites among pregnant people in the U.S are diphenyl phosphate (DPHP; parent compound, triphenyl phosphate (TPHP)) and bis(1,3-dichloro-2-propyl) phosphate (BDCIPP; parent compound, tris(1,3- dichloropropyl) (TDCIPP)), with greater than 95% detection frequencies in the National Health and Nutrition Examination Survey. 31,32 Growing experimental and observational evidence suggests that OPEs may affect early behavioral development at environmentally relevant doses via multiple biological mechanisms, including inflammation of various neuropathways, neurotransmitter perturbations, oxidative stress, and endocrine disruption. 33-37 Limited studies have reported associations between prenatal OPE exposures and increased externalizing behaviors, such as rule-breaking and aggression, and attention problems in children, with 162 most of those studies only examining the effects of the two most common OPE metabolites, BDCIPP and DPHP. 38,39 Similarly, studies have primarily examined the impacts of these single OPEs on neurobehavioral symptoms, rather than co-occurring impacts of multiple OPE exposures which are more representative of daily exposures mixtures. 38,40 A more thorough understanding of the impacts of prenatal OPE exposures on early neurobehavior is critical to developing appropriate interventions and regulations to mitigate neurotoxic exposures. This study’s objective was to evaluate the impacts of nine prenatal OPE metabolites alone and as a mixture on a broad range of early neurobehavioral symptoms, including internalizing, externalizing, and total problems, among mother-infant dyads participating in the MADRES cohort study. We hypothesized that higher prenatal exposures to OPE metabolites and OPE metabolite mixtures adversely impact child neurobehavioral outcomes at 36 months of age. METHODS Study Design The MADRES study is an ongoing prospective pregnancy cohort of predominately low-income Hispanic/Latino mother-child pairs living in urban Los Angeles, CA. A detailed description of the MADRES study population and protocol have been previously described. 41 In brief, participants were recruited into the study prior to 30 weeks’ gestation at three partner community health clinics, one private obstetrics and gynecology practice in Los Angeles, and through self-referrals from community meetings and local advertisements. Eligible participants at time of recruitment were: (1) less than 30 weeks’ gestation, (2) over 18 years of age, and (3) fluent in English or Spanish. Exclusion criteria included: (1) multiple gestation, (2) having a physical, mental, or cognitive disability that prevented participation or ability to provide consent, (3) current incarceration, and (4) HIV positive status. Written informed consent was obtained at study entry for each participant and the study was approved by the University of Southern California’s Institutional Review Board. 163 Nine urinary OPE metabolite concentrations were measured in 426 participants’ urine samples provided during the third trimester study visit (mean GA at sample collection ± SD= 31.4 ± 1.8 weeks) from 2017 to 2019. Child neurobehavior was assessed using the Child Behavioral Checklist (CBCL) composite scales, including the internalizing problems, externalizing problems, and total problems scales, administered at the 36-month timepoint. As shown in the consort diagram (Figure 1), mother-child participants with complete information on the exposure, outcome, and key covariates of interest were included in the final analytic sample. A total of 204 mother-child dyads with available data on OPE metabolite concentrations, the CBCL administered at 36 months, and key covariates were included in this study. OPE Metabolites Single spot urine samples were collected in 90 mL sterile specimen containers during a third trimester study visit. Urine specimens were aliquoted into 1.5 mL aliquot cryovials and specific gravity was measured in room temperature urine samples using a digital handheld refractometer (ATAGO PAL-10s pocket refractometer). Samples were stored at -80 ºC prior to shipment and sent to the Wadsworth Center’s Human Health Exposure Analysis Resource (HHEAR) lab hub for the analysis of the following nine OPE metabolites: diphenyl phosphate (DPHP), composite of di-n-butyl phosphate and di-isobutyl phosphate (DNBP + DIBP), bis(1,3,-dichloro-2-propyl) phosphate (BDCIPP), bis(2-chloroethyl) phosphate (BCEP), bis(butoxethyl) phosphate (BBOEP), bis(1-chloro-2-propyl) phosphate (BCIPP), bis(2-ethylhexyl) phosphate (BEHP), bis(2-methylphenyl) phosphate (BMPP), and dipropyl phosphate (DPRP). Additional information on each metabolite, the corresponding parent compound, and common uses are described in Chapter 1 (Table 1). Urinary OPE metabolites were quantified following methods similar to those previously described in the literature with some slight modifications and detailed in Chapter 2 (Prenatal Measures of Urinary OPE Metabolites). 42 In brief, high-performance liquid chromatography (HPLC, ExionLC™ system; SCIEX, Redwood City, CA, USA), coupled with an AB SCIEX QTRAP 5500+ triple quadrupole mass spectrometer (TQMS, Applied Biosystems, Foster City, CA, USA), was used in the identification and 164 quantification of target compounds. Target analytes limit of detection (LOD) ranged from 0.012 to 0.044 ng/mL. Due to poor chromatographic separation and co-elution of peaks accompanying a similar mass transition for DNBP and DIBP, these two isomers were reported as sum concentration of di-n-butyl phosphate and di-isobutyl phosphate (DNBP + DIBP). OPE metabolites with concentrations below the LOD were imputed using the LOD/√2. 43 Metabolites were then specific gravity (SG) adjusted using the following formula: Pc=P[(SGm-1)/(SG-1)], where Pc is the specific gravity corrected toxicant concentration (ng/mL), P is the observed toxicant concentration (ng/mL), SGm is the median SG value among the study population (median=1.016), and SG=the SG value of the sample. Health Outcome Assessment The Child Behavior Checklist for ages 1½ through 5 years old (CBCL 1.5-5) is a 99-item questionnaire which has been validated and widely used to assess a broad range of emotional and behavioral problems in children. 44 The questionnaire was orally administered to maternal participants during the 36 month study visit who indicated the frequency of behaviors in their child within the prior 2 months on a 3- point Likert scale (not true=0, sometimes true=1, or very often true=2), with each raw scale created by summing together relevant items and t-scores and corresponding borderline (T-scores: 60 – 63) and clinical symptom categories (T-scores: ≥64) calculated based on previously described criteria to quantify areas that may warrant evaluation by a professional. 45 Higher scores across all CBCL scales indicate increasing problems. The CBCL consists of seven scored syndrome scales (emotionally reactive (9 items), anxious/depressed (8 items), somatic complaints (11 items), withdrawn (8 items), sleep problems (7 items), attention problems (5 items), aggressive behavior (19 Items), and other problems (33 items)). These syndrome scales can be summed to create two composite scales, internalizing problems (emotionally reactive, anxious/depressed, somatic complaints, and withdrawn) and externalizing problems (attention problems and aggressive behavior). The CBCL additionally includes a total problems score which is the summed total of all 99 questionnaire items, plus the highest score on any additional problems listed under 165 an open-ended item, question 100 (score range=0-200). For the purposes of this analysis, the raw internalizing problems, externalizing problems, and total problems scores were each analyzed to encapsulate the breadth of potential behavioral and emotional developmental problems experienced by participants and to facilitate comparisons to prior studies similarly examining impacts of OPEs on raw CBCL scores. 40 However, sensitivity analyses examining associations between OPEs and CBCL T-scores were also evaluated to assess the robustness of our results after standardizing raw scores to a normative US sample of children. Covariates Covariates assessed in this analysis were study design or sample collection variables or were identified based on previous literature which examined impacts of neurotoxic chemicals on early neurobehavioral development. 3,31,39,40 Relationships between prenatal OPE metabolites and neurobehavioral development were visualized using a Directed Acyclic Graph (DAG) created using DAGitty (Figure S1). 46 All models were adjusted for variables identified in the DAG’s minimal sufficient adjustment set (maternal age, parity, pre-pregnancy BMI, race/ethnicity, income, and education) and study design or sample collection variables whose inclusion in models changed the effect estimate of our exposure of interest by 10% or more (recruitment site, specimen collection season, GA at sample collection, and child adjusted age at CBCL administration). The only exception to these criteria was adjustment for maternal-reported smoking during pregnancy. Prenatal smoking was identified in the minimal sufficient adjustment set, but, given the small frequency of maternal smoking (n=5, 2.5%), we instead evaluated its impact in sensitivity analyses by removing participants who reported smoking during pregnancy. Additionally, child sex was adjusted for in all models since it is an important predictor of neurobehavioral outcomes and was also evaluated as an effect modifier in adjusted models. Maternal age (years), household annual income during pregnancy (<$50,000, ≥ $50,000, do not know), education (≤12 th grade, >12 th grade), race/ethnicity (White non-Hispanic, Black non-Hispanic, Hispanic, Multiracial non-Hispanic/Other non-Hispanic), maternal smoking during pregnancy (yes, no), and parity (first born, ≥ second born, missing) were collected via interviewer administered questionnaires 166 in the participant’s preferred language (English or Spanish). Pre-pregnancy BMI was calculated using participant-reported pre-pregnancy weight and standing height measured by study staff at the first study visit using a commercial stadiometer (Perspectives Enterprises model P-AIM-101). Child sex assigned at birth was primarily abstracted from electronic medical records (n=200, 98.0%), followed by maternal- reported child sex (n=4, 2.0%) for cases in which abstracted sex could not be obtained. Statistical Analysis We examined participant demographic characteristics using means and frequencies. OPE metabolite distributions were explored using histograms, geometric means, percentile distributions, and metabolite detect frequencies. Given the generally right skewed distribution of OPE metabolites, Kruskal Wallis tests were conducted to evaluate bivariate associations between categorical covariates and OPE concentrations and Spearman correlations were performed to evaluate associations between OPE metabolites. The distribution of CBCL raw scores was right skewed with 7.4% and 2.5% of scores with a 0 on the internalizing and externalizing problems scales, respectively; therefore, CBCL scores were offset by 0.1 and natural log transformed prior to linear regression modeling. Locally Weighted Scatterplot Smoothing (LOWESS) plots between prenatal OPEs and CBCL composite scales were then evaluated, and due to non-linear associations that persisted after natural log transformation, OPE metabolites were categorized into exposure tertiles prior to linear regression modeling. For OPE biomarkers detected in >80% of participants (DPHP, DNBP+DIBP, BDCIPP), OPE metabolites were categorized into tertiles of specific gravity adjusted exposure concentrations. For OPE metabolites detected in 50-80% of participants (BCEP, BBOEP, BCIPP), a three-level categorical variable was created, with the lowest category defined as concentrations <LOD, and the remaining detected values categorized as <median or ≥median. For OPE biomarkers detected in <50% of participants (BMPP, BEHP, DPRP), we modeled OPE biomarkers as binary variables that were detected (>LOD) or not detected (≤LOD). Modeling assumptions for all linear regressions were evaluated and met. A statistical interaction between each OPE metabolite and child sex 167 was also tested in linear regression models. Data were managed and linear regression models were analyzed using SAS v9.4 (SAS Institute, Inc., Cary, NC, USA). Generalized Additive Models (GAMs) with a smoothing term for OPE metabolites were also performed to evaluate possible non-linear associations between OPE metabolites with a detect frequency >60% (DPHP, DNBP+DIBP, BDCIPP, BCEP, BBOEP) and neurobehavioral outcomes using the R package “mgcv.” A statistical interaction between each OPE metabolite and child sex was also tested within independent GAM models, using a factor smooth interaction. The significance level for single chemical analysis models was set at an alpha of 0.05. Bayesian kernel machine regression (BKMR) was selected as the primary mixture modeling approach given its ability to: 1) accommodate non-linear associations between an exposure and outcome of interest, while accounting for potential correlated exposures, and 2) evaluate possible synergistic and antagonistic relationships between mixtures components without prior specification. 47,48 Only metabolites with a detect frequency >60% were included in BKMR models (n=5 metabolites). BKMR is an advanced semi-parametric method which uses Gaussian kernel machine regression to estimate the effects of a high- dimensional matrix of predictors (e.g., interrelated environmental exposures) on a health outcome of interest. 47 The BKMR model for the current study is represented by the following equation: 𝑌 𝑖 = ℎ(𝐷𝑃𝐻𝑃 𝑖 , 𝐷𝑁𝐵𝑃 + 𝐷𝐼𝐵𝑃 𝑖 , 𝐵𝐷𝐶𝐼𝑃𝑃 𝑖 , 𝐵𝐶𝐸𝑃 𝑖 , 𝐵𝐵𝑂𝐸𝑃 𝑖 ) + 𝑋 𝑖 𝛽 + 𝜀 𝑖 where 𝑌 𝑖 represents our health outcome of interest (i.e., internalizing problems, externalizing problems, and total problems) for participant i, ℎ(.) denotes the exposure-response function; 𝛽 represents the vector of coefficients for model covariates (𝑋 𝑖 ), which are modeled parametrically; and 𝜀 represents residuals assumed to be independent, normally distributed, with a common variance. Five OPE metabolites detected in >60% of samples and CBCL raw composite scales were natural log transformed, mean-centered, and standard deviation scaled prior to BKMR modeling to facilitate comparisons. All continuous covariates were mean centered and scaled to one standard deviation. 168 The overall effect of the OPE mixture on each CBCL composite scale was evaluated by assessing the expected change in each score associated with concurrently increasing percentiles of all metabolites (DPHP, DNBP+DIBP, BDCIPP, BCEP, BBOEP), relative to fixing all metabolites at their median. If the 95% credible interval (CrI) did not span 0, we considered the metabolite or mixture to be associated with the outcome. Posterior inclusion probabilities (PIPs) were also estimated to assess the relative importance of each metabolite in the joint mixture effect with each CBCL composite raw score. Cross sections of the high-dimensional exposure-response functions were plotted for each OPE holding all other exposures constant at their 50 th percentiles to assess the shape, direction, and magnitude of association between each OPE metabolite, accounting for the rest of the mixture, with the CBCL composite scales. We also estimated the effect of an increase from the 25 th to the 75 th percentile of a single metabolite on each CBCL composite scale when all other metabolites were fixed at the median. Possible pairwise interactions between OPE metabolites were also investigated visually for each CBCL composite scale by assessing the association between each OPE metabolite and outcome when varying a second OPE metabolite to its 25 th , 50 th , and 75 th percentile (holding all other OPE metabolites at their 50 th percentile) with non-parallel lines indicating possible pairwise interactions. The bkmr R package (R v.4.1) was used for the BKMR analysis. 48 The Markov chain Monte Carlo (MCMC) sampler was used to obtain 100,000 posterior samples of model parameters, with the first half of iterations used as burn-in and chains thinned to every 10 th iteration to reduce potential autocorrelation. Visual inspection of trace plots and the Gelman-Rubin statistic were used to evaluate convergence, with both trace plots and Gelman-Rubin values below 1.1 indicating convergence. BKMR models were assumed to have non-informative prior distributions in primary models, the default specified in the R package. In order to further investigate possible synergistic and antagonistic relationships between OPE metabolites, a new Bayesian semiparametric regression was used to generate PIPs for interactions, using the NLinteraction R package. 49 This analysis was conducted by specifying 100,000 MCMC iterations, with half removed for burn-in and default options selected. The natural cubic splines used in this method to model the exposure-response relationship were based on the lowest value of Watanabe-Akaike information 169 criterion (WAIC), with the lowest WAIC for the internalizing model observed at 1 degree of freedom and the lowest WAIC for the total problems and externalizing models observed at 2 degrees of freedom. Pairwise interactions with the highest ranked PIPs were then further examined using GAMs, which allowed for a tensor interaction between the pair of metabolites (both evaluated continuously) and adjusted for the individual smoothed term of each metabolite and other covariates to obtain a p-value for the interaction. If interaction p-values were statistically significant (p<0.05) using GAMs, these relationships were further explored in models in which individual smoothed terms for one of the metabolites were assessed by tertiles of the second metabolite (and vice versa) to facilitate comparisons with the pairwise patterns observed in the BKMR analysis. Sensitivity Analysis Various sensitivity analyses were performed to assess the robustness of our results. Models excluding maternal participants who smoked during pregnancy were performed. An additional sensitivity analysis evaluating the effects of CBCL composite t-scores as an alternative parameterization of the CBCL raw scores was also performed. Since BKMR is sensitive to prior distributions, sensitivity analyses varying the parameter which controls the smoothness of the exposure-response association (b) were conducted; we explored both a lower (b=50) and higher (b=1000) degree of smoothness. Similarly, given NLinteraction’s sensitivity to model priors, we evaluated the impacts of varying the threshold parameter from the default (0.10) to a less conservative value (0.25). The threshold parameter influences the likelihood of metabolite inclusion into the function. RESULTS Descriptive Statistics The maternal and infant characteristics of participants analyzed in this study are shown in Table 1. Maternal participants were an average of 29.4±5.9 years old at study recruitment, had an average pre- pregnancy BMI of 29.1±6.5 kg/m 2 , and were predominately Hispanic (78.9%). Approximately half of participants completed at most high school (55.4%) and had an annual household income of less than 170 $50,000 during pregnancy (57.8%), and only 2.5% (n=5) of maternal participants reported smoking during pregnancy. Most infants were born full term, with an average gestational age at birth of 39.1±1.5 weeks. The distribution of OPE metabolite concentrations was similar in this analytical sample of participants compared with the full subset of 426 MADRES participants with prenatal OPE metabolites analyzed (Table S1). Similarly, this analytical subset was similar to both MADRES participants in the full cohort with children delivered during the study (n=774) and subset of MADRES participants with OPEs analyzed (n=426) for key demographic characteristics including income, ethnicity, maternal age, education, child sex, and infant GA at birth (see Table S2). Distributions of measured OPE metabolite concentrations are illustrated in Table 2. Median concentrations of BDCIPP (1.26 ng/mL) and DPHP (0.83 ng/mL) were higher than the other OPE metabolites investigated. Detection frequencies were greater than 60% for DPHP, DNBP+DIBP, BDCIPP, BCEP, and BBOEP and ranged between 26.0% and 51.5% for DPRP, BEHP, BMPP, and BCIPP. As shown in Figure 2, urinary OPE metabolites were weakly correlated with one another (Spearman 𝜌 = 0.01-0.27), with DPHP and BDCIPP having the highest correlation among all other OPE metabolites (Spearman 𝜌 =0.27). CBCL distributions among this sample of participants were approximately right skewed, with a median raw score of 6.0 (IQR: 9.0) for the internalizing problems scale, 8.0 (IQR: 12.0) for the externalizing problems scale, and 24.0 (IQR: 29.0) for the total problems scale (see Figure 3). Approximately 15.2% of participants had internalizing T-scores in the borderline to clinical range (borderline: 6.4%; clinical: 8.8%), 10.3% had externalizing T-scores in the borderline to clinical range (borderline: 4.9%; clinical: 5.4%), and 14.7% had total problems T-scores in the borderline to clinical range (borderline: 4.4%; clinical: 10.3%). Individual Metabolite Associations The unadjusted and adjusted associations between third trimester urinary OPE metabolites and children’s 36-month CBCL raw composite scores are shown in Table 3. Overall, higher concentrations of OPE metabolites were associated with higher externalizing, internalizing, and total problems scores in single metabolite adjusted models. When compared to non-detectable levels of BMPP metabolites, maternal 171 participants with detectable levels of BMPP exposure during the third trimester of pregnancy had children with significantly higher externalizing (𝛽 =1.39, 95% CI= 1.01, 1.90) and total problems scores (𝛽 =1.34, 95% CI= 1.02, 1.75) in unadjusted models, which remained statistically significant after adjustment for key covariates (externalizing 𝛽 =1.42, 95% CI= 1.04, 1.96; total problems 𝛽 =1.35, 95% CI= 1.03, 1.78). Maternal participants in the second tertile of BBOEP levels during pregnancy had children with higher externalizing scores (𝛽 =1.52, 95% CI= 1.05, 2.21) when compared to maternal participants in the first tertile of BBOEP in unadjusted models, which became marginally significant in adjusted models (𝛽 =1.43, 95% CI= 0.99, 2.09). However, children’s externalizing scores did not differ for levels in the third compared with first tertile of BBOEP, suggesting non-linear effects for this metabolite. There was a marginally significant association between maternal participants with detectable levels of BMPP concentrations and children with higher internalizing scores (𝛽 =1.45, 95% CI: 0.98, 2.14), relative to maternal participants with non-detectable BMPP levels. There were statistically significant interactions between prenatal levels of BCIPP and child sex for both internalizing scores (p=0.02) and total problems scores (p=0.03), and statistically significant associations in sex-stratified models. Among male children, internalizing scores (𝛽 = 2.20, 95% CI: 1.23, 3.95) and total problem scores (𝛽 =1.57, 95% CI: 1.06, 2.34) were higher for those with maternal metabolite levels in the third tertile of BCIPP compared with the first tertile. This association was not observed among female children between internalizing scores (𝛽 =0.61, 9%% CI: 0.30, 1.25) and total problems scores (𝛽 =0.69, 95% CI: 0.40, 1.17) and those with maternal metabolite levels in the third tertile of BCIPP, relative to the first tertile. We did not find any statistically significant associations for DPHP, DNBP+DIBP, BDCIPP, BCEP, BEHP, and DPRP with our three CBCL outcomes. However, the pattern of effects were generally suggestive of more linear patterns, with higher internalizing, externalizing, and total problems among children with mothers in the highest tertile of OPE metabolite concentrations compared with the lowest tertile. When compared to the linear regression model, we found evidence of a better model fit using GAMs for the association between prenatal BBOEP concentrations and children’s externalizing score at 36 172 months (p=0.04), with higher children’s externalizing scores at moderate concentrations of prenatal BBOEP but lower children’s externalizing scores at lower and higher concentrations of prenatal BBOEP (see Figure 4n). However, associations between prenatal urinary OPE metabolites and CBCL raw composite scores were not statistically significant when using GAMs, nor were interactions between each OPE metabolite and child sex. In sensitivity analyses evaluating CBCL T-scores as an alternative parametrization of CBCL scores, associations between maternal BMPP during pregnancy and children’s internalizing and externalizing T- scores were consistent with the associations observed when using the CBCL raw scores (see Table S3). Similar to our findings with raw scores, when compared to maternal participants in the first tertile of BBOEP exposure, maternal participants in the second tertile of BBOEP exposure had children with significantly higher externalizing (𝛽 =1.08, 95% CI= 1.00, 1.17) and total problem (𝛽 =1.09, 95% CI= 1.00, 1.18) T-scores, but results for the third tertile were not statistically significant across the internalizing, externalizing, or total problems scales. Results between OPE metabolites and CBCL composite T-scores in GAMs remained consistent with the patterns observed between OPE metabolites and CBCL composite raw scores (see Figure S2). In a sensitivity analysis excluding maternal participants who reported smoking during pregnancy, associations were consistent with the full study sample, both for linear regression models (see Table S4) and GAMs (see Figure S3). Mixtures Associations Concurrent increases in concentrations of all metabolites with CBCL composite raw scores had a non-monotonic, inverted U-shaped pattern, with lower CBCL composite scores at both higher and lower quantiles of metabolite mixtures when compared to the median. However, since all 95% CrI crossed 0, there were no cumulative associations between the overall OPE metabolite mixture and the internalizing, externalizing, and total problems raw scores (see Figure 5A, 5C, and 5E). Relationships between each individual metabolite, while fixing other metabolites at their median values, and children’s internalizing, externalizing, and total problems scores adjusting for key covariates are shown in Figure 5B, 5D, and 5F. A marginal association was observed between prenatal DNBP+DIBP 173 and the internalizing problems scale, with an increase in DNBP+DIBP from the 25 th to the 75 th percentile associated with a 0.15 (95% CrI: -0.02, 0.31) standard deviation increase on the internalizing problems scale, when all other metabolites were fixed at their median values and after adjustment for key covariates (Table 4). The association between BBOEP and each CBCL composite raw score was consistently non- linear and an inverted U-shaped, with higher internalizing, externalizing, and total problems scores among children at moderate concentrations of BBOEP but lower CBCL composite scores at lower and higher BBOEP concentrations. The associations between DNBP+DIBP and children’s total problems scores were positive and linear. However, the association between DNBP+DIBP and the externalizing score was relatively null. The shape and direction between BDCIPP, BCEP, and BBOEP and each CBCL composite raw score were consistent across scales; we observed an inverse, linear association with BDCIPP and each CBCL raw score and a positive and linear association between BCEP and each CBCL composite raw score. We found a relatively null association between DPHP and internalizing, externalizing, and total problems raw scores. Effect estimates evaluating the difference in CBCL composite raw scores for a change in the specified metabolite from the 25 th the 75th percentile, holding all other metabolites in the mixture at their median values and adjusting for key covariates, had 95% CrIs which spanned 0 (Table 4). Possible pairwise interactions between OPE metabolites and CBCL composite raw scores were visually identified using BKMR (Figure 6A, 6B, and 6C). PIPs for each pairwise interaction were also estimated using the NLinteraction method (Figure S4) and pairwise interactions with the highest ranked PIPs further examined. 49 In the internalizing scores model, the interaction between DNBP+DIBP and BCEP had the highest pairwise PIP estimated using NLinteraction (Figure S4). With BKMR, we observed a stronger positive association between DNBP+DIBP and internalizing scores at higher quartiles of BCEP. Within the externalizing scores model, the highest interaction PIP from NLinteraction was observed for DNBP+DIBP and BBOEP. With BKMR, we observed a positive association between DNBP+DIBP and externalizing scores at the 50 th and 75 th percentile of BBOEP, but an inverse association between DNBP+DIBP and externalizing scores at the 25 th percentile of BBOEP. In the total problems scores model, the largest interaction PIP identified by NLinteraction was for DNBP+DIBP and BCEP. With BKMR, we 174 observed a stronger positive association between DNBP+DIBP and total problems scores at higher quartiles of BCEP. Finally, the highest ranked interaction PIP for each CBCL composite score was further explored using GAMs to evaluate interaction p-values. We found a statistically significant interaction between DNBP+DIBP and total problems scores by BCEP concentrations modeled both continuously (p=0.03) and in tertiles (p=0.049), providing suggestive evidence of a potential interaction. Associations between prenatal DNBP+DIBP and children’s total problems scores by tertiles of BCEP were generally consistent with those observed in bivariate mixtures plots (see Figure S5). We found consistent results in sensitivity analyses evaluating CBCL T-scores as an alternative parametrization of the CBCL raw scores and as well as when we excluded mothers who reported smoking during pregnancy in our mixtures models (see Figure S6 and Figure S7). Results from models assuming a lower degree of smoothness (b=50) were very similar to primary results (Figure S8). However, the results for models assuming a higher degree of smoothness (b=1000) had a more linear pattern which appeared null in both the internalizing and total problems cumulative mixtures plots but maintained the inverted U- shaped pattern for the externalizing problems cumulative mixtures plots (Figure S9). Rankings of interaction PIPs from NLinteraction were consistent when increasing the value of the threshold parameter (Figure S10). DISCUSSION In this study of 204 predominately Hispanic and low-income mother-child dyads living in Los Angeles, California, we found important associations between independent OPE metabolites and neurobehavioral outcomes at 36 months of age as well as evidence for interacting effects between OPE metabolites and child’s sex. In single OPE analyses, detectable urinary BMPP concentrations during the third trimester of pregnancy were associated with higher internalizing problems, externalizing problems, and total problems in children at 36 months of age relative to those with non-detectable prenatal levels of BMPP. We also found that moderate (0.01-0.06 ng/mL) but not high (>0.06 ng/mL) concentrations of urinary BBOEP during pregnancy were associated with higher externalizing problems scores in children at 175 36 months of age when compared to those with non-detectable prenatal levels of BBOEP. Statistically significant non-linear and U-shaped patterns were observed between prenatal maternal BBOEP concentrations and children’s externalizing scores, with higher scores observed at moderate concentrations of BBOEP. Statistically significant interactions between BCIPP exposure and child’s sex were also identified for internalizing and total problems outcome models, with higher internalizing and total problems scores observed for male children whose mothers fell in the highest tertile of BCIPP compared with male children whose mothers fell in the lowest tertile of BCIPP. Although we did not observe an overall association between the mixture of DPHP, DNBP+DIBP, BDCIPP, BCEP, and BBOEP and neurobehavioral outcomes at 36-months, we did observe a positive association between prenatal DNBP+DIBP concentration and children’s internalizing problems, when fixing BDCIPP, BCEP, BBOEP and DPHP at their median concentrations. We also found evidence of a potential interaction between prenatal DNBP+DIBP and BCEP concentrations for total problems. Overall, our results suggest adverse effects of OPE exposures on neurobehavioral development, specifically for OPE metabolites commonly understudied and under monitored in pregnant individuals, with non-linear patterns and sex-specific interactions suggestive of endocrine-disrupting effects. Limited epidemiological evidence has reported adverse associations between prenatal OPE exposures and neurobehavioral outcomes in early childhood. The Pregnancy, Infection, and Nutrition (PIN) Study, a prospective birth cohort of predominately non-Hispanic white (~82%) and college educated individuals in North Carolina, found positive associations between BDCIPP and DPHP concentrations in prenatal urine and behavioral symptoms and externalizing problems using the Behavioral Assessment System for Children (BASC-2) among 199 children at 36 months of age. 38 The PIN study also reported an inverse association between isopropyl-phenyl phenyl phosphate (ip-PPP) and internalizing problems. The CHAMACOS study, a pregnancy cohort of predominately low-income and Hispanic participants in Central California found increased hyperactivity, using the BASC-2 at 7 years of age, with maternal urinary ip-PPP concentrations during pregnancy. 39 Another study by Choi et al., found a higher risk of ADHD among children with greater than median exposure to DPHP during pregnancy for participants in the Norwegian 176 mother, father, and child cohort study (MoBa), with more pronounced associations among girls and a decreased risk for ADHD with decreasing joint exposure to OPE metabolites, including DPHP and DNBP, and phthalates. 50 Other studies evaluating early life OPE exposures in dust concentrations on neurobehavioral outcomes have found similar adverse impacts between the summed exposure of OPEs (TDCPP, TPP, TCPP, and TCEP) and less responsible behavior and externalizing behavior problems using the teacher-rated Social Skills Improvement Rating Scale (SSIS). 51 Similarly, early exposures to TCEP in household dust have been associated with higher externalizing problems and early exposures to bisphenol A bis (diphenylphosphate) (BPA-BDPP) and resorcinol bis (diphenylphosphate) (PBDPP) in household dust have been associated with higher externalizing and internalizing problems at 18 months using the CBCL. 40 In our study, we did not observe statistically significant associations between BDCIPP and DPHP and externalizing symptoms, although the pattern for DPHP and externalizing symptoms in single metabolite models showed a similar direction of effect to prior literature. However, we observed adverse associations between detectable prenatal BMPP levels and higher internalizing, externalizing, and total problems and BBOEP concentrations and higher externalizing scores in single metabolite analyses. Additionally, positive associations between the highest tertile of BCIPP levels and male children’s internalizing and total problems scores were found. We also observed a marginal association between DNBP+DIBP and the internalizing problems scale when accounting for the rest of the mixture. Discrepancies in results across each of these studies may be attributable to a variety of factors, including but not limited to, heterogenous participant characteristics and exposure distributions, differences in the timing of exposure measurements (mid vs late gestation and varying years), outcome measurements, and children’s ages at behavioral assessments. For instance, the PIN study had higher median concentrations of DPHP (1.38 ng/mL vs. 0.83 ng/mL) and BDCIPP (2.01 ng/mL vs. 1.26 ng/mL) compared to MADRES participants; median concentrations among the CHAMACOS participants were relatively similar to those of MADRES for DPHP (0.9 ng/mL vs. 0.83 ng/mL) but lower for BDCIPP (0.4 ng/mL vs. 1.26 ng/mL). Participants in the MoBa cohort study had much lower median concentrations of DPHP (0.52 ng/mL vs. 177 0.83 ng/mL), BBOEP (0.08 ng/mL vs. 0.04), and BDCIPP (<LOD vs 1.26 ng/mL) compared to participants in the MADRES study. Additionally, the PIN study measured OPEs between 24-29 weeks’ gestation between 2001-2005, CHAMACOS at a mean gestational age of 26 weeks between 1999-2000, and MoBa at 17 weeks from 1999-2008, compared to MADRES at a mean gestational age of 31 weeks from 2017- 2019. The age at which children’s neurobehavioral development was assessed and the instruments used to measure neurobehavioral development also differed across these studies. While CHAMACOS assessed hyperactivity and attention problems using the BASC-2 when children were approximately 7 years old, the PIN study used all BASC-2 scales to evaluate neurobehavioral outcomes when children were 36 months of age. The MoBa study used the Norwegian Patient Registry to identify clinically diagnosed ADHD for children age 2.5 to 10 years. Despite these discrepancies across studies, the epidemiological literature generally suggests adverse impacts of OPE metabolites on neurobehavioral development. Emerging toxicological and epidemiological evidence suggests several mechanisms which may underlie the adverse association between prenatal exposures to environmentally relevant doses of OPEs and early behavioral and emotional development. Hypothesized mechanisms include direct impacts of prenatal OPEs on the neurological morphology and functioning of important neurobehavioral structures, including perturbations of glutamate and GABA neurotransmitters, 52-58 inflammation, 55,59 glia activation, 60,61 oxidative stress, 55,57,62 and decreased neuronal growth and network activity. 52,63-65 Furthermore, in animal studies using Wistar rats, the placenta has been implicated as a potentially important mechanism of developmental neurotoxicity from prenatal OPE exposures, with higher OPE accumulation in placental tissue among male placentas and further evidence of reduced forebrain serotonin (5-HT) and endocrine disruption, inflammation, and altered neurotransmitter production in the placenta. 66-69 Additional hypothesized mechanisms include maternal-mediated impacts of prenatal OPEs on early neurobehavior via critical mechanisms for neurobehavioral development, such as endocrine-disrupting pathways, which play a vital role in the development of the brain structures and processes important to behavior and which may be sex-specific. 70 Prior epidemiological studies have found an association between OPE exposures and altered levels of thyroid stimulating hormone (TSH) 71 and disruption of other thyroid hormones, 72 along 178 with disruption of sex-steroid hormones and sex-steroid binding globulins. 73 Given the rapid development of neurological systems during pregnancy, low-level chronic exposure to OPEs during pregnancy may exert neurotoxic effects on the developing fetus, with long-lasting neurobehavioral implications. 37,38 This study has several important strengths, beginning with its prospective design, which provided us with the opportunity to collect urine samples during potentially sensitive periods (i.e., pregnancy) to measure OPEs prior to our outcome of interest. An additional strength of this study was the use of prenatal urinary metabolites as a measure of in utero exposure to OPEs, given that maternal urinary OPE metabolites are considered reliable indicators of potential fetal OPE exposures. 15 We also measured various previously understudied OPE metabolites, including DNBP+DIBP, BCIPP, BCEP, BBOEP, DRPR, BMPP, and BEHP, which advances opportunities for risk assessment and subsequent interventions. Furthermore, the population evaluated in this study was largely comprised of pregnant individuals of Latin American ancestry, who are historically underrepresented in U.S. biomedical and population health research and disproportionally burdened by environmental exposures, 74 providing us with the opportunity to inform environmental justice solutions. An additional strength of this study is the use of a flexible environmental mixture modeling approach to assess the association between mixtures of OPE metabolites and neurobehavioral symptoms at 36 months. However, our study also has some limitations. Since single spot urine samples collected during the third trimester were used to assess OPE exposures throughout pregnancy, there may have been some exposure misclassification. However, previous studies indicate moderate to good reproducibility for OPE levels throughout pregnancy; although most have focused on DPHP and BDCIPP and fewer on many of the understudied OPE metabolites included in this study. 75,76 Additionally, although many key covariates identified in the literature were adjusted for, residual confounding could still be present, especially for postnatal OPE exposures, which could impact neurobehavioral outcomes. The small analytical sample analyzed in this study is another limitation since we may have been underpowered to detect associations between OPE mixtures and neurobehavioral outcomes. Furthermore, although our use of a flexible environmental mixture modeling approach was used to assess joint OPE exposures, we were unable to 179 explore the impacts of joint OPE exposures among metabolites with low detect frequencies, such as BMPP, which we found to adversely impact neurobehavioral development. CONCLUSION In this prospective pregnancy cohort of predominately low-income and Hispanic pregnant individuals living in Los Angeles, we found adverse associations between prenatal exposures to multiple previously understudied OPEs and children’s neurobehavioral symptoms at 36 months. There was also suggestive evidence of interactions between metabolites, highlighting the importance of evaluating OPEs beyond the effects of a single metabolite, along with non-linear and sex-specific associations between OPEs and children’s neurobehavioral development. Given the scarcity of studies evaluating associations between prenatal OPE metabolites and early neurobehavioral outcomes, additional studies exploring these associations, for exposures during both the prenatal and postnatal periods, are warranted. 180 Figure 1: Consort Diagram of Included Mother-Infant Dyads Mothers with prenatal OPE Urine Samples (n=426) Missing CBCL outcome at 3 years (n=41) Have not yet reached 3 years of age or passed the timepoint prior to its addition to the study (n=181) Complete OPE metabolites and Child Behavioral Checklist (CBCL) at 3 years of age (n=204) Missing data on key covariates (n=0) Final sample with complete data (n=204) 181 Figure 2: Spearman Correlations of Organophosphate Ester Metabolites (ng/mL) in Third Trimester Maternal Urine 182 Figure 3: Distributions of 36 Month Child Behavior Checklist Composite Raw Scores (CBCL) (N=204) 183 Figure 4: Associations Between Urinary Prenatal OPE Metabolite Concentrations (ng/mL) and CBCL Composite Raw Scores, Using Generalized Additive Models (N=204) Note: All models adjusted for recruitment site, maternal age, race/ethnicity, household annual income, education, pre-pregnancy BMI, GA at sample collection, child adjusted age at CBCL administration, season, infant birth order, child sex. OPE, Organophosphate Ester; CBCL, Child Behavior Check List; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. †Significant non-linearity † 184 Figure 5: Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and CBCL Composite Raw Scores, Using BKMR (n=204) Cumulative Mixture Univariate Plots Internalizing A. B. Externalizing C. D. Total Problems E. F. Figure 5 includes: 1) the estimated difference in CBCL composite score when setting all metabolites to the percentile specified on the x-axis compared with setting all metabolites to their median values (column 1), 2) the univariate relationship between each metabolite and CBCL outcome, while other metabolites are fixed at their medians, and a rug plot showing the distribution of the specified metabolite along the x-axis of each panel (column 2). All models were adjusted for recruitment site, maternal age, race/ethnicity, household annual income, education, pre-pregnancy BMI, GA at sample collection, child adjusted age at CBCL administration, season, infant birth order, child sex. OPE metabolites and CBCL raw scores were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were mean- centered and standard deviation scaled. Note: BKMR, Bayesian Kernel Machine Regression; OPE, Organophosphate Ester; CBCL, Child Behavior Check List; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 185 Figure 6: Bivariate Associations Between Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and CBCL Composite Raw Scores, Using BKMR (n=204) Bivariate Plots Internalizing A. Externalizing B. Total Problems C. Figure 6 shows the bivariate association between each OPE metabolite (labelled in the column) and CBCL composite score (Y axis), while setting a second metabolite (labelled in the row) to its 25 th , 50 th , and 75 th percentile and all other metabolites to their median. All models were adjusted for recruitment site, maternal age, race/ethnicity, household annual income, education, pre-pregnancy BMI, GA at sample collection, child adjusted age at CBCL administration, season, infant birth order, child sex. OPE metabolites and CBCL raw scores were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were mean-centered and standard deviation scaled. Note: BKMR, Bayesian Kernel Machine Regression; OPE, Organophosphate Ester; CBCL, Child Behavior Check List; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. Possible interactions were visually identified between the following metabolites for: internalizing scores (BDCIPP and BBOEP, DNBP+DIBP and BBOEP, DPHP and BBOEP, DNBP+DIBP and BCEP, DPHP and BCEP, BCEP and DNBP+DIBP, and DNBP+DIBP and BDCIPP), externalizing scores (BCEP and BBOEP, BDCIPP and BBOEP, DNBP+DIBP and BBOEP, and DPHP and BBOEP), and total problems scores (BCEP and BBOEP, BDCIPP and BBOEP, DNBP+DIBP and BBOEP, DPHP and BBOEP, DNBP+DIBP and BCEP, DPHP and BCEP, BCEP and DNBP+DIBP, BCEP and DPHP, and DNBP+DIBP and DPHP). 186 Table 1: Participant Characteristics (N=204) Mean (SD)/Freq(%) Maternal Characteristics Age (years) 29.4 (5.9) Education ≤High School ≥Technical school, college degree, or graduate studies 113 (55.4%) 91 (44.6%) Income Do not Know Less than $50,000 ≥ $50,000 55 (27.0%) 118 (57.8%) 31 (15.2%) NIH Race Categories White, non-Hispanic Black, non-Hispanic Hispanic Multiracial/other, non-Hispanic 17 (8.3%) 21 (10.3%) 161 (78.9%) 5 (2.5%) Smoking During Pregnancy No Yes 199 (98.0%) 5 (2.5%) Pre-pregnancy BMI (kg/m 2 ) 29.1 (6.5) Infant Characteristics Sex Female Male 105 (51.5%) 99 (48.5%) Infant Birth Order First Born Second or more Missing 74 (36.3%) 123 (60.3%) 7 (3.4%) Gestational Age at Birth (weeks) Child Adjusted Age at CBCL Administration (weeks) 39.1 (1.5) 155.8 (2.3) 187 Table 2. Distribution of Specific Gravity Adjusted OPE Concentrations (ng/mL) in Maternal Urine (N=204) Percentiles Distributions Metabolite 25th 50th 75th Min-Max Geometric Mean Detect Frequency LOD (ng/mL) DPHP 0.47 0.83 1.47 0.12-25.59 0.89 99.51% 0.0281 DNBP+DIBP 0.12 0.17 0.25 ND-1.78 0.18 96.57% 0.0441 BDCIPP 0.61 1.26 2.14 ND-34.94 1.05 95.10% 0.0174 BCEP 0.02 0.47 1.60 ND-168.00 0.31 68.63% 0.0200 BBOEP 0.02 0.04 0.08 ND-0.74 0.04 63.24% 0.0199 BCIPP ND 0.12 0.71 ND-19.90 0.11 51.47% 0.0204 BMPP ND 0.01 0.04 ND-0.47 0.02 39.71% 0.0115 BEHP ND ND 0.03 ND-3.48 0.03 25.00% 0.0170 DPRP ND ND 0.06 ND-2.85 0.04 25.98% 0.0278 Note: OPE, Organophosphate Esters; LOD, Limit of Detection; ND, Non-detect; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate; Min, Minimum; Max, Maximum. 188 Table 3: Individual Associations Between Third Trimester Urinary OPE Metabolites (ng/mL) and CBCL Raw Composite Scores (N=204) Internalizing Externalizing Total Problems Unadjusted 𝛽 (95% CI) Adjusted a 𝛽 (95% CI) Unadjusted 𝛽 (95% CI) Adjusted a 𝛽 (95% CI) Unadjusted 𝛽 (95% CI) Adjusted a 𝛽 (95% CI) DPHP T1 (<0.55) T2 (0.55-1.15) T3 (≥1.15) REF 0.82 (0.52, 1.31) 1.04 (0.65, 1.65) REF 0.89 (0.56, 1.41) 1.03 (0.64, 1.64) REF 0.86 (0.59, 1.26) 1.10 (0.75, 1.60) REF 0.87 (0.59, 1.26) 1.01 (0.68, 1.48) REF 0.85 (0.61, 1.17) 1.10 (0.79, 1.52) REF 0.87 (0.63, 1.20) 1.04 (0.75, 1.44) DNBP+DIBP T1 (<0.14) T2 (0.14-0.21) T3 (≥0.21) REF 1.01 (0.64, 1.61) 1.06 (0.66, 1.68) REF 1.06 (0.66, 1.69) 1.18 (0.74, 1.88) REF 1.09 (0.74, 1.60) 1.13 (0.77, 1.66) REF 1.04 (0.71, 1.52) 1.15 (0.78, 1.68) REF 1.00 (0.72, 1.38) 1.04 (0.75, 1.45) REF 0.97 (0.70, 1.35) 1.09 (0.78, 1.50) BDCIPP T1 (<0.85) T2 (0.85-1.83) T3 (≥1.83) REF 0.99 (0.62, 1.57) 0.86 (0.54, 1.37) REF 1.01 (0.62, 1.64) 0.97 (0.59, 1.61) REF 1.04 (0.71, 1.52) 1.15 (0.78, 1.68) REF 0.99 (0.66, 1.47) 1.16 (0.77, 1.75) REF 1.05 (0.75, 1.45) 1.07 (0.77, 1.48) REF 1.02 (0.72, 1.43) 1.12 (0.79, 1.59) BCEP T1 (<LOD) T2 (0.04-0.97) T3 (≥0.97) REF 1.10 (0.69, 1.75) 1.06 (0.66, 1.69) REF 1.19 (0.75, 1.90) 1.20 (0.74, 1.93) REF 1.06 (0.72, 1.55) 1.03 (0.70, 1.51) REF 1.14 (0.78, 1.68) 1.12 (0.75, 1.66) REF 1.00 (0.72) 1.04 (0.75, 1.44) REF 1.04 (0.75, 1.44) 1.09 (0.78, 1.53) BBOEP T1 (< LOD) T2 (0.01-0.06) T3 (≥0.06) REF 1.32 (0.83, 2.08) 0.88 (0.56, 1.38) REF 1.25 (0.79, 1.98) 0.78 (0.48, 1.25) REF 1.52 (1.05-2.21)* 0.97 (0.67, 1.41) REF 1.43 (0.99, 2.09) 0.87 (0.59, 1.29) REF 1.34 (0.97, 1.84) 0.93 (0.67, 1.27) REF 1.26 (0.92, 1.74) 0.84 (0.60, 1.17) BCIPP T1 (<LOD) T2 (0.03- 0.66) T3 (≥0.66) REF 0.78 (0.49, 1.23) 1.32 (0.84, 2.09) REF 0.72 (0.46, 1.15) 1.47 (0.93, 2.34) REF 0.90 (0.62, 1.32) 1.36 (0.93, 1.98) REF 0.89 (0.61, 1.31) 1.33 (0.99, 1.95) REF 0.87 (0.63, 1.20) 1.22 (0.88, 1.68) REF 0.85 (0.62, 1.18) 1.21 (0.87, 1.67) BMPP Non-detect Detect REF 1.42 (0.97, 2.08) REF 1.45 (0.98, 2.14) REF 1.39 (1.01, 1.90)* REF 1.42 (1.04, 1.96)* REF 1.34 (1.02, 1.75)* REF 1.35 (1.03, 1.78)* BEHP Non-detect Detect REF 1.21 (0.78, 1.87) REF 1.13 (0.73, 1.76) REF 1.08 (0.75, 1.55) REF 1.01 (0.77, 1.45) REF 1.15 (0.85, 1.56) REF 1.10 (0.81, 1.50) DPRP Non-detect Detect REF 0.84 (0.55, 1.29) REF 0.85 (0.56, 1.31) REF 1.08 (0.75, 1.53) REF 1.10 (0.77, 1.56) REF 0.96 (0.71, 1.30) REF 0.97 (0.72, 1.31) a Model adjusted for recruitment site, maternal age, race/ethnicity, household annual income, education, pre-pregnancy BMI, GA at sample collection, child adjusted age at CBCL administration, season, infant birth order, child sex. Note: OPE, Organophosphate Ester; T1, Tertile 1; T2, Tertile 2; T3, Tertile 3; <LOD, Below Limit of Detection; CBCL, Child Behavior Check List; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) 189 phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate; GA, gestational age; BMI, Body Mass Index. All 𝛽 ′𝑠 have been exponentiated for interpretation. 190 Table 4: Posterior Inclusion Probabilities (PIPs) and Single Exposure Effect Estimates for Each Prenatal OPE Metabolite in the Bayesian Kernel Machine Regression (BKMR) Mixture and CBCL Composite Raw Score Metabolite PIPs Effect Estimates 95% Credible Interval Internalizing Scale DPHP 0.27 0.002 -0.17, 0.18 DNBP+DIBP 0.53 0.15 -0.02, 0.32 BDCIPP 0.34 -0.09 -0.24, 0.06 BCEP 0.37 0.15 -0.10, 0.41 BBOEP 0.64 a -0.10 -0.33, 0.14 Externalizing Scale DPHP 0.15 0.04 -0.13, 0.20 DNBP+DIBP 0.19 0.02 -0.13, 0.17 BDCIPP 0.15 -0.04 -0.17, 0.10 BCEP 0.20 0.11 -0.14, 0.36 BBOEP 0.81 a -0.02 -0.28, 0.24 Total Problems Scale DPHP 0.24 0.03 -0.14, 0.21 DNBP+DIBP 0.30 0.06 -0.10, 0.22 BDCIPP 0.26 -0.04 -0.19, 0.10 BCEP 0.28 0.11 -0.15, 0.36 BBOEP 0.59 a -0.07 -0.31, 0.17 a Highest PIPs value. Effect estimates reflect the difference in CBCL composite score for a change in the specified metabolite from the 25th to 75th percentile, holding all other metabolites in the mixture at their median values and adjusting for recruitment site, maternal age, race/ethnicity, household annual income, education, pre-pregnancy BMI, GA at sample collection, child adjusted age at CBCL administration, season, infant birth order, child sex. Note: OPE, Organophosphate Ester; CBCL, Child Behavior Check List; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 191 SUPPLEMENTAL FIGURES Supplemental Figure 1: Directed Acyclic Graph (DAG) of Prenatal OPE metabolites and Child Neurobehavioral Development Directed Acyclic Graph (DAG) used to identify potential confounders and precision variables. The DAG was created using DAGitty. Green ovals represent exposures or predictors of the exposure, pink ovals represent potential confounders, and blue ovals represent the outcome or predictors of outcome. Minimally sufficient set: Maternal age, parity, pre-pregnancy BMI, prenatal smoking, race/ethnicity, socioeconomic status (SES) 193 Supplemental Figure 2: Associations Between Urinary Prenatal OPE Metabolite Concentrations (ng/mL) and CBCL Composite T-Scores, Using Generalized Additive Models (N=204) All models adjusted for recruitment site, maternal age, race/ethnicity, household annual income, education, pre- pregnancy BMI, GA at sample collection, child adjusted age at CBCL administration, season, infant birth order, child sex. Note: OPE, Organophosphate Ester; CBCL, Child Behavior Check List; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. †Significant non-linearity † † † 194 Supplemental Figure 3: Associations Between Urinary Prenatal OPE Metabolite Concentrations (ng/mL) and CBCL Composite Raw Scores Among Participants Who Reported No In-Utero Smoking, Using Generalized Additive Models (n=199) All models adjusted for recruitment site, maternal age, race/ethnicity, household annual income, education, pre- pregnancy BMI, GA at sample collection, child adjusted age at CBCL administration, season, infant birth order, child sex. Note: OPE, Organophosphate Ester; CBCL, Child Behavior Check List; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 195 Supplemental Figure 4: Posterior Inclusion Probabilities (PIPs) for Pairwise Interactions Between OPE Metabolites and CBCL Composite Raw Scores Using NLinteraction Method Internalizing Externalizing Total Problems A. B. C. Posterior inclusion probabilities for each of the five metabolites were estimated using the NLinteraction method. The light blue color reflects higher PIPs scores across metabolites. All models were adjusted for recruitment site, maternal age, race/ethnicity, household annual income, education, pre-pregnancy BMI, GA at sample collection, child adjusted age at CBCL administration, season, infant birth order, child sex. OPE metabolites and CBCL raw composite scores were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were mean-centered and standard deviation scaled. Note: BKMR, Bayesian Kernel Machine Regression; OPE, Organophosphate Ester; CBCL, Child Behavior Check List; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 196 Supplemental Figure 5: Prenatal DNBP+DIBP Exposures and Children’s Total Problems Scores by Tertiles of BCEP, Using Generalized Additive Models Note: DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BCEP, Bis(2- chloroethyl) phosphate. 197 Supplemental Figure 6: Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and CBCL Composite T- Scores, Using BKMR (N=204) Cumulative Mixture Univariate Plots Bivariate Plots Internalizing A. B. C. Externalizing D. E. F. Total Problems G. H. I. Figure 6 includes: 1) the estimated difference in CBCL composite score when setting all metabolites to the percentile specified on the x-axis compared with setting all metabolites to their median values (column 1), 2) the univariate relationship between each metabolite and CBCL outcome, while other metabolites are fixed at their medians, and a rug plot showing the distribution of the specified metabolite along the x-axis of each panel (column 2), and 3) the bivariate association between each OPE metabolite (labelled in the column) and CBCL composite score (Y axis), while setting a second metabolite (labelled in the row) to its 25 th , 50 th , and 75 th percentile and all other metabolites to their median. All models were adjusted for recruitment site, maternal age, race/ethnicity, household annual income, education, pre-pregnancy BMI, GA at sample collection, child adjusted age at CBCL administration, season, infant birth order, child sex. OPE metabolites were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were mean-centered and standard deviation scaled. Note: BKMR, Bayesian Kernel Machine Regression; OPE, Organophosphate Ester; CBCL, Child Behavior Check List; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 198 Supplemental Figure 7: Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and CBCL Composite Raw Scores Among Participants Who Reported No In-Utero Smoking, Using BKMR (N=199) Cumulative Mixture Univariate Plots Bivariate Plots Internalizing A. B. C. Externalizing D. E. F. Total Problems G. H. I. Figure 7 includes: 1) the estimated difference in CBCL composite score when setting all metabolites to the percentile specified on the x-axis compared with setting all metabolites to their median values (column 1), 2) the univariate relationship between each metabolite and CBCL outcome, while other metabolites are fixed at their medians, and a rug plot showing the distribution of the specified metabolite along the x-axis of each panel (column 2), and 3) the bivariate association between each OPE metabolite (labelled in the column) and CBCL composite score (Y axis), while setting a second metabolite (labelled in the row) to its 25 th , 50 th , and 75 th percentile and all other metabolites to their median (column 3). All models were adjusted for recruitment site, maternal age, race/ethnicity, household annual income, education, pre-pregnancy BMI, GA at sample collection, child adjusted age at CBCL administration, season, infant birth order, child sex. OPE metabolites and CBCL raw composite scores were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were mean-centered and standard deviation scaled. Note: BKMR, Bayesian Kernel Machine Regression; OPE, Organophosphate Ester; CBCL, Child Behavior Check List; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 199 Supplemental Figure 8: Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and CBCL Composite Raw Scores, Using BKMR Varying the Smoothing Parameter to b=50 Cumulative Mixture Univariate Plots Bivariate Plots Internalizing A. B. C. Externalizing D. E. F. Total Problems G. H. I. Figure 8 includes: 1) the estimated difference in CBCL composite score when setting all metabolites to the percentile specified on the x-axis compared with setting all metabolites to their median values (column 1), 2) the univariate relationship between each metabolite and CBCL outcome, while other metabolites are fixed at their medians, and a rug plot showing the distribution of the specified metabolite along the x-axis of each panel (column 2), and 3) the bivariate association between each OPE metabolite (labelled in the column) and CBCL composite score (Y axis), while setting a second metabolite (labelled in the row) to its 25 th , 50 th , and 75 th percentile and all other metabolites to their median. All models were adjusted for recruitment site, maternal age, race/ethnicity, household annual income, education, pre-pregnancy BMI, GA at sample collection, child adjusted age at CBCL administration, season, infant birth order, child sex. OPE metabolites and CBCL raw composite scores were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were mean-centered and standard deviation scaled. Note: BKMR, Bayesian Kernel Machine Regression; OPE, Organophosphate Ester; CBCL, Child Behavior Check List; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 200 Supplemental Figure 9: Prenatal OPE Urinary Metabolite Mixtures (ng/mL) and CBCL Composite Raw Scores, Using BKMR Varying the Smoothing Parameter to b=1000 Cumulative Mixture Univariate Plots Bivariate Plots Internalizing A. B. C. Externalizing D. E. F. Total Problems G. H. I. Figure 9 includes: 1) the estimated difference in CBCL composite score when setting all metabolites to the percentile specified on the x-axis compared with setting all metabolites to their median values (column 1), 2) the univariate relationship between each metabolite and CBCL outcome, while other metabolites are fixed at their medians, and a rug plot showing the distribution of the specified metabolite along the x-axis of each panel (column 2), and 3) the bivariate association between each OPE metabolite (labelled in the column) and CBCL composite score (Y axis), while setting a second metabolite (labelled in the row) to its 25 th , 50 th , and 75 th percentile and all other metabolites to their median. All models were adjusted for recruitment site, maternal age, race/ethnicity, household annual income, education, pre-pregnancy BMI, GA at sample collection, child adjusted age at CBCL administration, season, infant birth order, child sex. OPE metabolites and CBCL raw composite scores were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were mean-centered and standard deviation scaled. Note: BKMR, Bayesian Kernel Machine Regression; OPE, Organophosphate Ester; CBCL, Child Behavior Check List; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 201 Supplemental Figure 10: Posterior Inclusion Probabilities (PIPs) for Pairwise Interactions Between OPE Metabolites and CBCL Composite Raw Scores Using NLinteraction Method and Increasing the Threshold to 0.25 Internalizing Externalizing Total Problems A. B. C. Posterior inclusion probabilities for each of the five metabolites were estimated using the NLinteraction method. The light blue color reflects higher PIPs scores across metabolites. All models were adjusted for recruitment site, maternal age, race/ethnicity, household annual income, education, pre-pregnancy BMI, GA at sample collection, child adjusted age at CBCL administration, season, infant birth order, child sex. OPE metabolites and CBCL raw composite scores were natural log-transformed, mean centered, and standard deviation scaled. Continuous covariates were mean-centered and standard deviation scaled. Note: BKMR, Bayesian Kernel Machine Regression; OPE, Organophosphate Ester; CBCL, Child Behavior Check List; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis (2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate. 202 Supplemental Table 1. Distribution of Specific Gravity Adjusted OPE Concentrations (ng/mL) in Urine for Maternal Participants Analyzed (N=204) vs the Full Sample of Maternal Participants with OPEs Available (N=426) Subset Analyzed (N=204) Available OPEs (N=426) Metabolite 25th 50th 75th Min-Max Detect Frequency 25th 50th 75th Min-Max Detect Frequency LOD (ng/mL) DPHP 0.47 0.83 1.47 0.12-25.59 99.51% 0.47 0.77 1.46 0.12-25.59 99.77% 0.0281 DNBP+DIBP 0.12 0.17 0.25 ND-1.78 96.57% 0.12 0.18 0.26 ND-4.29 97.65% 0.0441 BDCIPP 0.61 1.26 2.14 ND-34.94 95.10% 0.61 1.29 2.29 ND-68.00 94.60% 0.0174 BCEP 0.02 0.47 1.60 ND-168.00 68.63% 0.02 0.53 1.62 ND-168.00 68.31% 0.0200 BBOEP 0.02 0.04 0.08 ND-0.74 63.24% 0.02 0.04 0.07 ND- 1.17 64.79% 0.0199 BCIPP ND 0.12 0.71 ND-19.90 51.47% ND 0.18 0.77 ND-40.56 53.76% 0.0204 BMPP ND 0.01 0.04 ND-0.47 39.71% ND 0.01 0.04 ND- 0.69 38.73% 0.0115 BEHP ND ND 0.03 ND-3.48 25.00% ND ND 0.04 ND-4.42 25.59% 0.0170 DPRP ND ND 0.06 ND-2.85 25.98% ND ND 0.05 ND-2.85 24.18% 0.0278 Note: OPE, Organophosphate Esters; LOD, Limit of Detection; ND, Non-detect; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2- ethylhexyl) phosphate; DPRP, Dipropyl phosphate; Min, Minimum; Max, Maximum. 203 Supplemental Table 2: Comparison of Participant Characteristics Analyzed in the Analytical Dataset (N=204) to Subset with OPE Metabolite Concentrations Available (N=426) and Full MADRES Participants who Have Delivered Children in the Study (N=774) Mean (SD)/Freq(%) Mean (SD)/Freq(%) Mean (SD)/Freq(%) Maternal Characteristics N=204 N=426 N=774 Age (years) 29.4 (5.9) 28.9 (6.1) 28.3 (6.0) Education ≤High School ≥High School Missing 113 (55.4%) 91 (44.6%) ─ 240 (56.3%) 182 (42.7%) 4 (0.9%) 416 (53.7%) 316 (40.8%) 42 (5.4%) Income Don’t Know Less than $50,000 ≥ $50,000 Missing 55 (27.0%) 118 (57.8%) 31 (15.2%) ─ 122 (28.6%) 249 (58.5%) 51 (12.0%) 4 (0.9%) 244 (31.5%) 416 (53.7%) 73 (9.4%) 41 (5.3%) NIH Race Categories White, non-Hispanic Black, non-Hispanic Hispanic Multiracial/other, non-Hispanic Missing 17 (8.3%) 21 (10.3%) 161 (78.9%) 5 (2.5%) ─ 29 (6.8%) 49 (11.5%) 329 (77.2%) 15 (3.5%) 4 (0.9%) 40 (5.2%) 87 (11.2%) 579 (74.8%) 26 (3.4%) 42 (5.4%) Smoking During Pregnancy No Yes Missing 199 (98.0%) 5 (2.5%) ─ 418 (98.1%) 8 (1.9%) ─ 659 (85.1%) 13 (1.7%) 102 (13.2%) Pre-pregnancy BMI (kg 2 /m) 29.1 (6.5) 28.6 (6.7) 28.7 (6.7) Infant Characteristics Sex Female Male Missing 105 (51.5%) 99 (48.5%) ─ 218 (51.2%) 208 (48.8) ─ 384 (49.6%) 387 (50.0%) 3 (0.4%) Infant Birth Order First Born Second or more Missing 74 (36.3%) 123 (60.3%) 7 (3.4%) 147 (34.5%) 261 (61.3%) 18 (4.2) 242 (31.3%) 406 (52.5%) 126 (16.3%) Gestational Age at Birth (weeks) Infant Adjusted Age (weeks) 39.1 (1.5) 155.8 (2.3) 39.1 (1.5) ─ 39.0 (1.8) ─ 205 Supplemental Table 3: Individual Associations Between Third Trimester Urinary OPE Metabolites (ng/mL) and CBCL Composite T-Scores (n=204) Internalizing Externalizing Total Problems Unadjusted 𝛽 (95% CI) Adjusted a 𝛽 (95% CI) Unadjusted 𝛽 (95% CI) Adjusted a 𝛽 (95% CI) Unadjusted 𝛽 (95% CI) Adjusted a 𝛽 (95% CI) DPHP T1 (<0.55) T2 (0.55-1.15) T3 (≥1.15) REF 0.99 (0.91, 1.07) 1.03 (0.94, 1.12) REF 1.00 (0.92, 1.09) 1.03 (0.94, 1.12) REF 1.00 (0.93, 1.09) 1.04 (0.97, 1.13) REF 1.00 (0.93, 1.09) 1.03 (0.95, 1.11) REF 0.99 (0.92, 1.08) 1.04 (0.96, 1.13) REF 1.00 (0.92, 1.09) 1.03 (0.94, 1.12) DNBP+DIBP T1 (<0.14) T2 (0.14-0.21) T3 (≥0.21) REF 0.99 (0.91, 1.07) 1.00 (0.92, 1.09) REF 0.99 (0.91, 1.08) 1.01 (0.93, 1.10) REF 1.03 (0.95, 1.11) 1.03 (0.96, 1.12) REF 1.02 (0.94, 1.11) 1.03 (0.96, 1.12) REF 1.01 (0.93, 1.10) 1.01 (0.93, 1.09) REF 1.01 (0.93, 1.09) 1.01 (0.93, 1.10) BDCIPP T1 (<0.85) T2 (0.85-1.83) T3 (≥1.83) REF 1.01 (0.92, 1.09) 0.98 (0.90, 1.07) REF 1.02 (0.93, 1.11) 1.01 (0.92, 1.11) REF 1.02 (0.94, 1.10) 1.03 (0.95, 1.11) REF 1.01 (0.93, 1.10) 1.03 (0.95, 1.13) REF 1.01 (0.93, 1.10) 1.02 (0.94, 1.10) REF 1.01 (0.92, 1.10) 1.03 (0.94, 1.12) BCEP T1 (<LOD) T2 (0.04-0.97) T3 (≥0.97) REF 1.02 (0.94, 1.11) 1.01 (0.93, 1.10) REF 1.03 (0.95, 1.12) 1.03 (0.94, 1.13) REF 0.99 (0.92, 1.07) 0.99 (0.91, 1.07) REF 1.01 (0.93, 1.09) 1.01 (0.93, 1.09) REF 1.00 (0.92, 1.09) 1.00 (0.92, 1.08) REF 1.01 (0.93, 1.10) 1.01 (0.93, 1.10) BBOEP T1 (< LOD) T2 (0.01-0.06) T3 (≥0.06) REF 1.07 (0.99, 1.17) 1.01 (0.93, 1.09) REF 1.07 (0.98, 1.16) 0.99 (0.91, 1.08) REF 1.09 (1.01, 1.18) 1.01 (0.94, 1.09) REF 1.08 (1.00, 1.17) 0.99 (0.92, 1.08) REF 1.10 (1.01, 1.19) 1.01 (0.93, 1.10) REF 1.09 (1.00, 1.18) 1.00 (0.91, 1.08) BCIPP T1 (<LOD) T2 (0.03- 0.66) T3 (≥0.66) REF 0.96 (0.88, 1.04) 1.04 (0.96, 1.13) REF 0.95 (0.87, 1.03) 1.05 (0.96, 1.14) REF 0.99 (0.91, 1.07) 1.05 (0.97, 1.14) REF 0.98 (0.91, 1.06) 1.04 (0.96, 1.13) REF 0.97 (0.90, 1.06) 1.04 (0.96, 1.13) REF 0.97 (0.89, 1.06) 1.04 (0.96, 1.13) BMPP Non-detect Detect REF 1.08 (1.01, 1.16) REF 1.08 (1.01, 1.16) REF 1.06 (0.99, 1.13) REF 1.07 (1.00, 1.14) REF 1.08 (1.01, 1.15) REF 1.08 (1.01, 1.16) BEHP Non-detect Detect REF 1.04 (0.96, 1.13) REF 1.03 (0.95, 1.12) REF 1.02 (0.95, 1.10) REF 1.01 (0.94, 1.08) REF 1.03 (0.96, 1.12) REF 1.02 (0.94, 1.10) DPRP Non-detect Detect REF 0.99 (0.92, 1.07) REF 1.00 (0.92, 1.08) REF 1.03 (0.96, 1.11) REF 1.03 (0.96, 1.11) REF 1.02 (0.94, 1.10) REF 1.02 (0.95, 1.10) a Model adjusted for recruitment site, maternal age, race/ethnicity, household annual income, education, pre-pregnancy BMI, GA at sample collection, child adjusted age at CBCL administration, season, infant birth order, child sex. 206 Note: OPE, Organophosphate Ester; T1, Tertile 1; T2, Tertile 2; T3, Tertile 3; <LOD, Below Limit of Detection; CBCL, Child Behavior Check List; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate; GA, gestational age; BMI, Body Mass Index. All 𝛽 ′𝑠 have been exponentiated to facilitate interpretation. 207 Supplemental Table 4: Individual Associations Between Third Trimester Urinary OPE Metabolites (ng/mL) and CBCL Raw Composite Scores Among Participants Who Reported No In-Utero Smoking (N= 199) Internalizing 𝛽 (95% CI) Externalizing 𝛽 (95% CI) Total Problems 𝛽 (95% CI) DPHP T1 (<0.55) T2 (0.55-1.15) T3 (≥1.15) REF 0.88 (0.55, 1.41) 1.01 (0.62, 1.62) REF 0.84 (0.58, 1.23) 0.98 (0.66, 1.44) REF 0.86 (0.62, 1.18) 1.01 (0.72, 1.41) DNBP+DIBP T1 (<0.14) T2 (0.14-0.21) T3 (≥0.21) REF 1.05 (0.65, 1.70) 1.17 (0.73, 1.88) REF 1.01 (0.68, 1.48) 1.11 (0.76, 1.64) REF 0.95 (0.68, 1.33) 1.06 (0.76, 1.48) BDCIPP T1 (<0.85) T2 (0.85-1.83) T3 (≥1.83) REF 1.00 (0.61, 1.65) 1.00 (0.59, 1.67) REF 1.01 (0.67, 1.51) 1.24 (0.82, 1.89) REF 1.03 (0.73, 1.46) 1.18 (0.82, 1.68) BCEP T1 (<LOD) T2 (0.04-0.97) T3 (≥0.97) REF 1.18 (0.74, 1.90) 1.15 (0.70, 1.89) REF 1.13 (0.77, 1.66) 1.04 (0.70, 1.55) REF 1.03 (0.74, 1.43) 1.04 (0.74, 1.46) BBOEP T1 (< LOD) T2 (0.01-0.06) T3 (≥0.06) REF 1.22 (0.76, 1.94) 0.74 (0.45, 1.20) REF 1.37 (0.94, 1.99) 0.79 (0.54, 1.17) REF 1.22 (0.88, 1.68) 0.78 (0.56, 1.09) BCIPP T1 (<LOD) T2 (0.03- 0.66) T3 (≥0.66) REF 0.73 (0.45, 1.17) 1.49 (0.93, 2.39) REF 0.90 (0.61, 1.32) 1.35 (0.92, 1.98) REF 0.86 (0.62, 1.19) 1.22 (0.88, 1.70) BMPP Non-detect Detect REF 1.49 (1.00, 2.21) REF 1.46 (1.06, 2.01) REF 1.38 (2.05, 1.81) BEHP Non-detect Detect REF 1.14 (0.72, 1.79) REF 1.01 (0.70, 1.46) REF 1.10 (0.81, 1.51) DPRP Non-detect Detect REF 0.87 (0.56, 1.35) REF 1.12 (0.79, 1.60) REF 0.98 (0.73, 1.33) All models adjusted for recruitment site, maternal age, race/ethnicity, household annual income, education, pre-pregnancy BMI, GA at sample collection, child adjusted age at CBCL administration, season, infant birth order, child sex. Note: OPE, Organophosphate Ester; T1, Tertile 1; T2, Tertile 2; T3, Tertile 3; <LOD, Below Limit of Detection; CBCL, Child Behavior Check List; DPHP, Diphenyl phosphate; DNBP+DIBP, Sum of Dibutyl phosphate and Di-isobutyl phosphate; BDCIPP, Bis(1,3-dichloro-2-propyl) phosphate; BCEP, Bis(2- chloroethyl) phosphate; BBOEP, Bis(butoxethyl) phosphate; BCIPP, Bis(1-chloro-2-propyl) phosphate; BMPP, Bis(2-methylphenyl) phosphate; BEHP, Bis(2-ethylhexyl) phosphate; DPRP, Dipropyl phosphate; GA, gestational age; BMI, Body Mass Index. All 𝛽 ′𝑠 have been exponentiated to facilitate interpretation. 208 REFERENCES 1. Salisbury AL, Fallone MD and Lester B. Neurobehavioral assessment from fetus to infant: the NICU Network Neurobehavioral Scale and the Fetal Neurobehavior Coding Scale. Mental retardation and developmental disabilities research reviews 2005; 11: 14-20. DOI: 10.1002/mrdd.20058. 2. Monk C and Hane A. Fetal and Infant Brain–Behavior Development: Milestones & Environmental Influences. 2014. 3. Grandjean P and Landrigan PJ. Neurobehavioural effects of developmental toxicity. The Lancet Neurology 2014; 13: 330-338. DOI: 10.1016/S1474-4422(13)70278-3. 4. Rauh VA and Margolis AE. Research Review: Environmental exposures, neurodevelopment, and child mental health - new paradigms for the study of brain and behavioral effects. J Child Psychol Psychiatry 2016; 57: 775-793. 2016/03/19. DOI: 10.1111/jcpp.12537. 5. Barker DJP. The origins of the developmental origins theory. Journal of Internal Medicine 2007; 261: 412-417. https://doi.org/10.1111/j.1365-2796.2007.01809.x. DOI: https://doi.org/10.1111/j.1365- 2796.2007.01809.x. 6. Pantelaki I and Voutsa D. Organophosphate flame retardants (OPFRs): A review on analytical methods and occurrence in wastewater and aquatic environment. Sci Total Environ 2019; 649: 247-263. 2018/09/03. DOI: 10.1016/j.scitotenv.2018.08.286. 7. Yang J, Zhao Y, Li M, et al. A Review of a Class of Emerging Contaminants: The Classification, Distribution, Intensity of Consumption, Synthesis Routes, Environmental Effects and Expectation of Pollution Abatement to Organophosphate Flame Retardants (OPFRs). Int J Mol Sci 2019; 20 2019/06/20. DOI: 10.3390/ijms20122874. 8. Blum A, Behl M, Birnbaum L, et al. Organophosphate Ester Flame Retardants: Are They a Regrettable Substitution for Polybrominated Diphenyl Ethers? Environ Sci Technol Lett 2019; 6: 638- 649. 2020/06/05. DOI: 10.1021/acs.estlett.9b00582. 9. Dodson RE, Perovich LJ, Covaci A, et al. After the PBDE Phase-Out: A Broad Suite of Flame Retardants in Repeat House Dust Samples from California. Environmental Science & Technology 2012; 46: 13056-13066. DOI: 10.1021/es303879n. 10. Bergman A, Rydén A, Law RJ, et al. A novel abbreviation standard for organobromine, organochlorine and organophosphorus flame retardants and some characteristics of the chemicals. Environ Int 2012; 49: 57-82. 2012/09/13. DOI: 10.1016/j.envint.2012.08.003. 11. Stapleton HM, Klosterhaus S, Eagle S, et al. Detection of organophosphate flame retardants in furniture foam and U.S. house dust. Environ Sci Technol 2009; 43: 7490-7495. 2009/10/24. DOI: 10.1021/es9014019. 12. Stapleton HM, Klosterhaus S, Keller A, et al. Identification of flame retardants in polyurethane foam collected from baby products. Environ Sci Technol 2011; 45: 5323-5331. 2011/05/20. DOI: 10.1021/es2007462. 13. Hoffman K, Butt CM, Webster TF, et al. Temporal Trends in Exposure to Organophosphate Flame Retardants in the United States. Environmental Science & Technology Letters 2017; 4: 112-118. DOI: 10.1021/acs.estlett.6b00475. 14. Vuong AM, Yolton K, Cecil KM, et al. Flame retardants and neurodevelopment: An updated review of epidemiological literature. Curr Epidemiol Rep 2020; 7: 220-236. 2021/01/08. DOI: 10.1007/s40471-020-00256-z. 15. Hou R, Xu Y and Wang Z. Review of OPFRs in animals and humans: Absorption, bioaccumulation, metabolism, and internal exposure research. Chemosphere 2016; 153: 78-90. DOI: https://doi.org/10.1016/j.chemosphere.2016.03.003. 16. Bollmann UE, Möller A, Xie Z, et al. Occurrence and fate of organophosphorus flame retardants and plasticizers in coastal and marine surface waters. Water research 2012; 46: 531-538. 209 17. Dodson RE, Rodgers KM, Carey G, et al. Flame retardant chemicals in college dormitories: flammability standards influence dust concentrations. Environmental science & technology 2017; 51: 4860-4869. 18. Giulivo M, Capri E, Kalogianni E, et al. Occurrence of halogenated and organophosphate flame retardants in sediment and fish samples from three European river basins. Science of the Total Environment 2017; 586: 782-791. 19. Ali N, Malik RN, Mehdi T, et al. Organohalogenated contaminants (OHCs) in the serum and hair of pet cats and dogs: biosentinels of indoor pollution. Science of the total environment 2013; 449: 29-36. 20. Yang F, Ding J, Huang W, et al. Particle size-specific distributions and preliminary exposure assessments of organophosphate flame retardants in office air particulate matter. Environmental science & technology 2014; 48: 63-70. 21. Mö ller A, Sturm R, Xie Z, et al. Organophosphorus flame retardants and plasticizers in airborne particles over the Northern Pacific and Indian Ocean toward the polar regions: Evidence for global occurrence. Environmental science & technology 2012; 46: 3127-3134. 22. Kim J-W, Isobe T, Chang K-H, et al. Levels and distribution of organophosphorus flame retardants and plasticizers in fishes from Manila Bay, the Philippines. Environmental pollution 2011; 159: 3653-3659. 23. Bika SH, Adeniji AO, Okoh AI, et al. Spatiotemporal Distribution and Analysis of Organophosphate Flame Retardants in the Environmental Systems: A Review. Molecules 2022; 27: 573. DOI: 10.3390/molecules27020573. 24. National Academies of Sciences E and Medicine. Why Indoor Chemistry Matters. Washington, DC: The National Academies Press, 2022, p.190. 25. Varshavsky JR, Robinson JF, Zhou Y, et al. Organophosphate Flame Retardants, Highly Fluorinated Chemicals, and Biomarkers of Placental Development and Disease During Mid-Gestation. Toxicol Sci 2021; 181: 215-228. 2021/03/08. DOI: 10.1093/toxsci/kfab028. 26. Ding J, Xu Z, Huang W, et al. Organophosphate ester flame retardants and plasticizers in human placenta in Eastern China. Sci Total Environ 2016; 554-555: 211-217. 2016/03/08. DOI: 10.1016/j.scitotenv.2016.02.171. 27. Zhao F, Chen M, Gao F, et al. Organophosphorus Flame Retardants in Pregnant Women and Their Transfer to Chorionic Villi. Environ Sci Technol 2017; 51: 6489-6497. 2017/05/19. DOI: 10.1021/acs.est.7b01122. 28. Bai XY, Lu SY, Xie L, et al. A pilot study of metabolites of organophosphorus flame retardants in paired maternal urine and amniotic fluid samples: potential exposure risks of tributyl phosphate to pregnant women. Environ Sci Process Impacts 2019; 21: 124-132. 2018/11/15. DOI: 10.1039/c8em00389k. 29. Doherty BT, Hammel SC, Daniels JL, et al. Organophosphate Esters: Are These Flame Retardants and Plasticizers Affecting Children's Health? Curr Environ Health Rep 2019; 6: 201-213. 2019/11/23. DOI: 10.1007/s40572-019-00258-0. 30. Wang X, Chen P, Zhao L, et al. Transplacental Behaviors of Organophosphate Tri- and Diesters Based on Paired Human Maternal and Cord Whole Blood: Efficiencies and Impact Factors. Environ Sci Technol 2021; 55: 3091-3100. 2021/01/06. DOI: 10.1021/acs.est.0c06095. 31. Hoffman K, Lorenzo A, Butt CM, et al. Predictors of urinary flame retardant concentration among pregnant women. Environ Int 2017; 98: 96-101. 2016/10/13. DOI: 10.1016/j.envint.2016.10.007. 32. Ospina M, Jayatilaka NK, Wong LY, et al. Exposure to organophosphate flame retardant chemicals in the U.S. general population: Data from the 2013-2014 National Health and Nutrition Examination Survey. Environ Int 2018; 110: 32-41. 2017/11/06. DOI: 10.1016/j.envint.2017.10.001. 33. Dishaw LV, Powers CM, Ryde IT, et al. Is the PentaBDE replacement, tris (1,3-dichloro-2- propyl) phosphate (TDCPP), a developmental neurotoxicant? Studies in PC12 cells. Toxicol Appl Pharmacol 2011; 256: 281-289. 2011/01/19. DOI: 10.1016/j.taap.2011.01.005. 210 34. Program NT. NTP Toxicology and Carcinogenesis Studies of Tris (2-chloroethyl) Phosphate (CAS No. 115-96-8) in F344/N Rats and B6C3F1 Mice (Gavage Studies). National Toxicology Program technical report series 1991; 391: 1-233. 35. Tilson H, Veronesi B, McLamb R, et al. Acute exposure to tris (2-chloroethyl) phosphate produces hippocampal neuronal loss and impairs learning in rats. Toxicol Appl Pharmacol 1990; 106: 254-269. 36. Wang Q, Lai NL-S, Wang X, et al. Bioconcentration and Transfer of the Organophorous Flame Retardant 1,3-Dichloro-2-propyl Phosphate Causes Thyroid Endocrine Disruption and Developmental Neurotoxicity in Zebrafish Larvae. Environmental Science & Technology 2015; 49: 5123-5132. DOI: 10.1021/acs.est.5b00558. 37. Patisaul HB, Behl M, Birnbaum LS, et al. Beyond Cholinesterase Inhibition: Developmental Neurotoxicity of Organophosphate Ester Flame Retardants and Plasticizers. Environ Health Perspect 2021; 129: 105001. 2021/10/07. DOI: 10.1289/ehp9285. 38. Doherty BT, Hoffman K, Keil AP, et al. Prenatal exposure to organophosphate esters and behavioral development in young children in the Pregnancy, Infection, and Nutrition Study. Neurotoxicology 2019; 73: 150-160. 2019/04/06. DOI: 10.1016/j.neuro.2019.03.007. 39. Castorina R, Bradman A, Stapleton HM, et al. Current-use flame retardants: Maternal exposure and neurodevelopment in children of the CHAMACOS cohort. Chemosphere 2017; 189: 574-580. 2017/10/01. DOI: 10.1016/j.chemosphere.2017.09.037. 40. Sugeng EJ, de Cock M, Leonards PEG, et al. Association of exposure to organophosphate flame retardants and children's behavior at a median age of 18 months. Environmental Advances 2021; 5: 100077. DOI: https://doi.org/10.1016/j.envadv.2021.100077. 41. Bastain TM, Chavez T, Habre R, et al. Study Design, Protocol and Profile of the Maternal And Developmental Risks from Environmental and Social Stressors (MADRES) Pregnancy Cohort: a Prospective Cohort Study in Predominantly Low-Income Hispanic Women in Urban Los Angeles. BMC Pregnancy Childbirth 2019; 19: 189-189. DOI: 10.1186/s12884-019-2330-7. 42. Wang Y, Li W, Martínez-Moral MP, et al. Metabolites of organophosphate esters in urine from the United States: Concentrations, temporal variability, and exposure assessment. Environ Int 2019; 122: 213-221. 2018/11/20. DOI: 10.1016/j.envint.2018.11.007. 43. Hornung RW and Reed LD. Estimation of Average Concentration in the Presence of Nondetectable Values. Applied Occupational and Environmental Hygiene 1990; 5: 46-51. DOI: 10.1080/1047322X.1990.10389587. 44. Achenbach TM. Manual for the ASEBA preschool forms & profiles : an integrated system of multi-informant assessment. In: 2000. 45. Achenbach. Manual for the ASEBA preschool forms & profiles: an integrated system of multi- informant assessment. . 2000. 46. Textor J, van der Zander B, Gilthorpe MS, et al. Robust causal inference using directed acyclic graphs: the R package 'dagitty'. Int J Epidemiol 2016; 45: 1887-1894. 2017/01/17. DOI: 10.1093/ije/dyw341. 47. Bobb JF, Valeri L, Claus Henn B, et al. Bayesian kernel machine regression for estimating the health effects of multi-pollutant mixtures. Biostatistics 2015; 16: 493-508. 2014/12/24. DOI: 10.1093/biostatistics/kxu058. 48. Bobb JF, Claus Henn B, Valeri L, et al. Statistical software for analyzing the health effects of multiple concurrent exposures via Bayesian kernel machine regression. Environmental Health 2018; 17: 67. DOI: 10.1186/s12940-018-0413-y. 49. Antonelli J, Mazumdar M, Bellinger D, et al. Estimating the health effects of environmental mixtures using Bayesian semiparametric regression and sparsity inducing priors. The Annals of Applied Statistics 2020; 14: 257-275. 50. Choi G, Keil AP, Richardson DB, et al. Pregnancy exposure to organophosphate esters and the risk of attention-deficit hyperactivity disorder in the Norwegian mother, father and child cohort study. Environ Int 2021; 154: 106549. DOI: https://doi.org/10.1016/j.envint.2021.106549. 211 51. Lipscomb ST, McClelland MM, MacDonald M, et al. Cross-sectional study of social behaviors in preschool children and exposure to flame retardants. Environ Health 2017; 16: 23. 2017/03/10. DOI: 10.1186/s12940-017-0224-6. 52. Hausherr V, van Thriel C, Krug A, et al. Impairment of glutamate signaling in mouse central nervous system neurons in vitro by tri-ortho-cresyl phosphate at noncytotoxic concentrations. Toxicol Sci 2014; 142: 274-284. 2014/09/10. DOI: 10.1093/toxsci/kfu174. 53. Hogberg HT, de Cássia da Silveira ESR, Kleensang A, et al. Organophosphorus flame retardants are developmental neurotoxicants in a rat primary brainsphere in vitro model. Arch Toxicol 2021; 95: 207-228. 2020/10/21. DOI: 10.1007/s00204-020-02903-2. 54. Yang W, Zhao F, Fang Y, et al. (1)H-nuclear magnetic resonance metabolomics revealing the intrinsic relationships between neurochemical alterations and neurobehavioral and neuropathological abnormalities in rats exposed to tris(2-chloroethyl)phosphate. Chemosphere 2018; 200: 649-659. 2018/03/09. DOI: 10.1016/j.chemosphere.2018.02.056. 55. Liu X, Zhao X, Wang Y, et al. Triphenyl phosphate permeates the blood brain barrier and induces neurotoxicity in mouse brain. Chemosphere 2020; 252: 126470. 2020/05/24. DOI: 10.1016/j.chemosphere.2020.126470. 56. Gant DB, Eldefrawi ME and Eldefrawi AT. Action of organophosphates on GABAA receptor and voltage-dependent chloride channels. Fundam Appl Toxicol 1987; 9: 698-704. 1987/11/01. DOI: 10.1016/0272-0590(87)90176-x. 57. Wang Q, Lai NL, Wang X, et al. Bioconcentration and transfer of the organophorous flame retardant 1,3-dichloro-2-propyl phosphate causes thyroid endocrine disruption and developmental neurotoxicity in zebrafish larvae. Environ Sci Technol 2015; 49: 5123-5132. 2015/04/01. DOI: 10.1021/acs.est.5b00558. 58. Shi Q, Wang M, Shi F, et al. Developmental neurotoxicity of triphenyl phosphate in zebrafish larvae. Aquat Toxicol 2018; 203: 80-87. 2018/08/11. DOI: 10.1016/j.aquatox.2018.08.001. 59. Zhong X, Wu J, Ke W, et al. Neonatal exposure to organophosphorus flame retardant TDCPP elicits neurotoxicity in mouse hippocampus via microglia-mediated inflammation in vivo and in vitro. Archives of Toxicology 2020; 94: 541-552. DOI: 10.1007/s00204-019-02635-y. 60. Slotkin TA, Skavicus S, Stapleton HM, et al. Brominated and organophosphate flame retardants target different neurodevelopmental stages, characterized with embryonic neural stem cells and neuronotypic PC12 cells. Toxicology 2017; 390: 32-42. DOI: https://doi.org/10.1016/j.tox.2017.08.009. 61. Hogberg HT, de Cássia da Silveira E Sá R, Kleensang A, et al. Organophosphorus flame retardants are developmental neurotoxicants in a rat primary brainsphere in vitro model. Archives of Toxicology 2021; 95: 207-228. DOI: 10.1007/s00204-020-02903-2. 62. Wu S, Ji G, Liu J, et al. TBBPA induces developmental toxicity, oxidative stress, and apoptosis in embryos and zebrafish larvae (Danio rerio). Environmental Toxicology 2016; 31: 1241-1249. https://doi.org/10.1002/tox.22131. DOI: https://doi.org/10.1002/tox.22131. 63. Behl M, Hsieh J-H, Shafer TJ, et al. Use of alternative assays to identify and prioritize organophosphorus flame retardants for potential developmental and neurotoxicity. Neurotoxicology and Teratology 2015; 52: 181-193. DOI: https://doi.org/10.1016/j.ntt.2015.09.003. 64. Ryan KR, Sirenko O, Parham F, et al. Neurite outgrowth in human induced pluripotent stem cell- derived neurons as a high-throughput screen for developmental neurotoxicity or neurotoxicity. NeuroToxicology 2016; 53: 271-281. DOI: https://doi.org/10.1016/j.neuro.2016.02.003. 65. Shafer TJ, Brown JP, Lynch B, et al. Evaluation of Chemical Effects on Network Formation in Cortical Neurons Grown on Microelectrode Arrays. Toxicological Sciences 2019; 169: 436-455. DOI: 10.1093/toxsci/kfz052. 66. Phillips AL, Chen A, Rock KD, et al. Editor’s highlight: transplacental and lactational transfer of Firemaster® 550 components in dosed Wistar rats. Toxicological Sciences 2016; 153: 246-257. 67. Baldwin KR, Phillips AL, Horman B, et al. Sex specific placental accumulation and behavioral effects of developmental Firemaster 550 exposure in Wistar rats. Scientific reports 2017; 7: 1-13. 212 68. Rock KD, Horman B, Phillips AL, et al. EDC IMPACT: Molecular effects of developmental FM 550 exposure in Wistar rat placenta and fetal forebrain. Endocrine connections 2018; 7: 305. 69. Rock KD, St Armour G, Horman B, et al. Effects of prenatal exposure to a mixture of organophosphate flame retardants on placental gene expression and serotonergic innervation in the fetal rat brain. Toxicological Sciences 2020; 176: 203-223. 70. Ghassabian A, Henrichs J and Tiemeier H. Impact of mild thyroid hormone deficiency in pregnancy on cognitive function in children: lessons from the Generation R Study. Best practice & research Clinical endocrinology & metabolism 2014; 28: 221-232. 71. Yao Y, Li M, Pan L, et al. Exposure to organophosphate ester flame retardants and plasticizers during pregnancy: Thyroid endocrine disruption and mediation role of oxidative stress. Environ Int 2021; 146: 106215. DOI: https://doi.org/10.1016/j.envint.2020.106215. 72. Demeneix Barbara A. Evidence for Prenatal Exposure to Thyroid Disruptors and Adverse Effects on Brain Development. European Thyroid Journal 2019; 8: 283-292. DOI: 10.1159/000504668. 73. Luo K, Liu J, Wang Y, et al. Associations between organophosphate esters and sex hormones among 6-19-year old children and adolescents in NHANES 2013-2014. Environ Int 2020; 136: 105461. 2020/01/14. DOI: 10.1016/j.envint.2020.105461. 74. Buckley JP, Kuiper JR, Bennett DH, et al. Exposure to Contemporary and Emerging Chemicals in Commerce among Pregnant Women in the United States: The Environmental influences on Child Health Outcome (ECHO) Program. Environmental Science & Technology 2022; 56: 6560-6573. DOI: 10.1021/acs.est.1c08942. 75. Hoffman K, Daniels JL and Stapleton HM. Urinary metabolites of organophosphate flame retardants and their variability in pregnant women. Environ Int 2014; 63: 169-172. 2013/12/10. DOI: 10.1016/j.envint.2013.11.013. 76. Romano ME, Hawley NL, Eliot M, et al. Variability and predictors of urinary concentrations of organophosphate flame retardant metabolites among pregnant women in Rhode Island. Environmental Health 2017; 16: 1-11. 213 CHAPTER 6 SUMMARY Neurodevelopment occurs across the lifespan and is continuously modified by a complex host of environmental, psychosocial, and biological influences which can impact the neurological pathways responsible for healthy brain functioning. 1 The prenatal period is increasingly recognized as a critical period of early neurodevelopment which is highly susceptible to exogenous influences, with any inhibitions or alterations to neurological processes having the potential to impact integral structures important for healthy neurodevelopment. 2 As a result, environmental pollutants, including industrial chemicals, pose a serious risk of neurotoxicity and are of increasing public health concern, especially considering the high prevalence of neurodevelopmental disorders and their costly social, economic, and health effects. 3-5 Since industrial chemicals are modifiable risk factors, early identification of their potential neurotoxicity provides us with intervenable opportunities. OPE flame retardants have increased in use in the past decade, as previous historical (polychlorinated biphenyls (PCBs)) and legacy (PBDEs) flame retardants were phased out due to concerns over their neurotoxic effects on children. 6,7 Although originally intended to prevent and delay fires in the hopes of protecting against physical harm and the destruction of property, emerging evidence suggests adverse impacts from OPEs on reproductive outcomes, preterm birth and children’s respiratory outcomes, measures of adiposity, and neurodevelopment. 8,9 However, the epidemiological literature on the effects of prenatal OPEs on children’s neurodevelopment is limited, with even less evidence on the impacts of simultaneous exposures to multiple OPEs. The current investigation aimed to contribute epidemiological evidence on the prenatal effects of OPE concentrations on early neurodevelopmental outcomes, with the evaluation of many previously understudied OPEs, including BCEP, BBOEP, BCIPP, BEHP, and DPRP. Since chemical exposures do not occur in isolation and simultaneous exposure to multiple OPEs is likely, another important aim of this investigation was to understand how exposures to multiple OPEs impacted neurodevelopment and how 214 results varied from single pollutant models evaluating impacts of OPEs on neurodevelopment. We further evaluated these associations among a predominantly Hispanic and low-income population, which is disproportionally burdened by environmental exposures, and which has historically been underrepresented in the U.S. biomedical and population health research. 10,11 This provides us with the opportunity to inform environmental justice solutions. In the first study of this dissertation (Sex-Specific Effects of Prenatal Organophosphate Ester (OPE) Metabolite Mixtures and Adverse Infant Birth Outcomes in the Maternal and Developmental Risks from Environmental and Social Stressors (MADRES) Pregnancy Cohort) 12 our primary aim was to examine the association between prenatal urinary OPE metabolites, independently and in mixtures, on gestational age at birth and birthweight. We found that early GA at birth was associated with independent prenatal urinary OPE metabolites in sex-specific ways, with earlier GA at birth associated with higher BDCIPP concentrations among male infants and DNBP+DIBP among female infants. The overall effect of the urinary OPE mixture was also associated with earlier GA at birth among all infants, but more pronounced effects were observed among female infants. We did not observe any associations between prenatal urinary OPE metabolites and BW-for-GA z scores. The second study of this dissertation (Prenatal Exposures to Organophosphate Ester Metabolites and Early Motor Development in the MADRES Cohort) examined associations between prenatal urinary OPEs metabolites and gross and fine motor development throughout infancy (6, 9, 12, and 18 months) among 329 participants in the MADRES cohort. We found that a doubling in DPHP concentrations was associated with a 26% increased risk of motor delays across infancy and an 11% decrease in expected fine motor scores among infants. Additionally, we observed sex-specific associations between prenatal DPRP and gross motor development and BCIPP and fine motor development, with adverse impacts on female infants’ motor development. However, detectable prenatal BMPP levels were associated with increased gross motor scores among infants’, although the association was not observed using instrument-suggested cut-offs for risk of motor delay. 215 In the third study of this dissertation (Prenatal Exposures to Organophosphate Ester Metabolite Mixtures and Children’s Neurobehavioral Outcomes in the MADRES Pregnancy Cohort), we evaluated the association between prenatal OPE metabolites, independently and in mixtures, on neurobehavioral symptoms among 204 36-month-old children. In single metabolite models, we found associations between detectable levels of BMPP during pregnancy and higher externalizing and total problems. Additionally, participants in the third tertile of prenatal BBOEP concentrations had higher externalizing scores, relative to the first tertile; however, the associations were not statistically significant. Although the OPE metabolite mixture and each CBCL composite outcome were not statistically significantly associated, we observed a marginal association between DNBP+DIBP and higher internalizing scores, when holding other metabolites at their median. IMPLICATIONS FOR POLICY AND PRACTICE Overall, this dissertation found evidence of adverse developmental impacts from higher concentrations of prenatal OPEs, with adverse effects observed from birth to early childhood. Consistent with prior toxicological and epidemiological evidence, 13-16 we observed sex-specific effects from prenatal OPE metabolites across gestational age at birth, gross and fine motor development, and neurobehavioral symptoms, implicating the endocrine system as a potentially important mechanism underlying OPEs adverse effects on neurodevelopment. Our work additionally highlighted the importance of considering the impacts of multiple OPE exposures during pregnancy on children’s health outcomes, as a more comprehensive picture of the impacts of prenatal OPEs on children’s health was continuously observed when using mixtures approaches when compared to single pollutant models. Alongside previous epidemiological evidence, our findings contribute to the growing literature implicating OPEs as neurotoxicants. 8,16-19 Despite existing limitations in our understanding of OPEs neurotoxic effects on children, there is considerable evidence to suggest that global reduction efforts of OPEs may be warranted to protect children’s health, especially considering that potential delays in exposure reduction efforts could have adverse impacts on children’s health during developmentally susceptible periods. 20 However, it is important to acknowledge that flame retardants play an important role in fire 216 prevention within our society, delaying and/or suppressing fires to minimize the potentially disastrous destruction of property and loss of human life which accompany a fire. 21 As a result, any potential regulatory phase outs or mitigation efforts in the absence of flame-retardant alternatives and an integrated plan or essentiality assessment moving forward could be disastrous, possibly creating fire hazards or resulting in regrettable substitutions, as has previously occurred for flame retardants. 22 We have seen similar patterns of regrettable substitutions with other harmful chemicals, such as Bisphenol A (BPA), which were subsequently replaced with harmful replacements, like bisphenol S (BPS). 23 Flame retardants have a similar history of regrettable substitutions, with OPEs originally intended to replace PBDEs, which were intended to replace PCBs. 22 Thus, flame retardants may serve as an important case study of the challenges experienced during two previous chemical phase outs, providing us with the opportunity to recognize patterns which contributed to the repetitive cycle of regrettable chemical substitutions and allowing us to pose some potential solutions to those challenges. However, it is crucial to acknowledge that this is a highly complicated issue which will likely require an integrated and interdisciplinary plan, spearheaded by a variety of policy, engineering, and scientific experts, along with multi-stakeholder collaborations, to address comprehensively and adequately. The following case study is but a modest attempt to consider some of the previous challenges experienced during two prior phase outs of chemical flame retardants and pose potential recommendations but is not intended to be an exhaustive list. FLAME RETARDANTS AND NEUROTOXICITY: A CASE STUDY ON A REPETITIVE CYCLE OF REGRETTABLE SUBSTITUTIONS Chemical flame retardants (FRs) are a diverse group of chemical additives incorporated into a variety of consumer products to delay or suppress fires. 24 Generally classified by their active chemical element, FRs suppress fires in a variety of mechanisms, including interfering with a fire’s ability to consume oxygen, forming a barrier, or acting as a chemical coolant. 21 Chemical flame retardants are traditionally used by manufacturers to meet flammability regulations, particularly California’s Technical Bulletin 117 (TB-117), which implemented a multitude of open flame tests and resilient filling material requirements in 1972 for upholstered furniture, in response to a growing number of house fires in which furniture was often 217 involved. 6,21 Since 1972, TB-117 has undergone several revisions (TB117-2013) to remove some of the more stringent open flame tests and added additional laws (SB1019) requiring the clear use of labeling when selling chemically treated flame-retardant furniture. 6 Along with flammability regulations which have encouraged the use of chemical flame retardants in upholstered furniture, chemical flame retardants are commonly used in electronic materials, building and construction materials, and transportation vehicles such as airplanes, cars, and trains, to meet fire safety standards. 25 Since chemical flame retardants are commonly applied as additives, they can easily leach from consumer products into the air and attach to air particles and then dust, foods, and water. 9,26 Flame retardants may also contaminate the air, water, or soil during the manufacturing process or during the dismantling or burning of electronic materials. 27 As a result, humans are chronically exposed to chemical flame-retardants, with common exposure routes including dermal, inhalation, and ingestion which, given evidence of adverse effects on human health, has recently become a cause for concern, particularly among susceptible populations like children. 8,9 During the 1970s, polybrominated biphenyls (PBBs) and polychlorinated biphenyls (PCBs) were commonly used flame-retardant chemicals due to their fire resistance properties. 28 However, PBBs and PCBs were phased out of most industrialized countries in 1976 as evidence of their toxicity, harm to human health, and bioaccumulation emerged. 28 For instance, prenatal exposures to PCBs concentrations were associated with adverse neurological outcomes, with cognitive impairments and impacts on verbal memory among adolescents and behavioral issues, such as increased abnormal behavior in childhood, ADHD symptoms, and longer reaction times, among children. 29-34 After PBBs and PCBs were phased out, polybrominated diphenyl ethers (PBDEs) increased in use in the late 1970s until their subsequent phase out in 2003 due to similar concerns regarding their toxicity to human health and bioaccumulation. 6 Prenatal and postnatal concentrations of PBDEs have been associated with adverse impacts on full scale IQ in children, along with positive associations between prenatal PBDE concentrations and increased externalizing problems. 35-37 As PBDEs were phased out, OPEs increased in production volume in the US, rapidly becoming the most widely used flame retardant. 6,7,38 However, emerging toxicological and epidemiological evidence suggests potential neurotoxicity concerns, 20,22 with prenatal and postnatal OPEs 218 associated with adverse cognitive and behavioral problems. 8,17-19,39 In response, some OPEs, specifically TDCIPP, have been added to CA Proposition 65, and new reporting regulations have emerged, such as Washington Children’s Safe Product Act, requiring chemical reporting. This rather persistent cycle of regrettable chemical flame-retardant substitutions potentially stemmed from a variety of factors. For instance, when PBBs and PCBs were phased out in 1976, PBDEs had already been in use as flame retardants since the 1960s. 22 Due to their structural similarity to PCBs and PBBs, they easily replaced their use during the manufacturing process, filling a chemical flame retardancy vacuum via substitution to a similar chemical. 40 There was additionally a lack of data on the potential health impacts of PBDEs, so a transition to a new chemical with a similar function occurred easily, especially considering the functional need to meet flammability regulations. 40 When PBDEs were ultimately phased out and their status as regrettable substitutions widely acknowledged, there was again the creation of a similar vacuum for chemical flame retardants. 23 Since OPEs have half-lives of a couple of hours to days, they were originally assumed to be less persistent than PBDEs and as a result, less harmful. 6 Similar to the transition from PCBs/PBBs to PBDEs, there was a lack of data on the potential health impacts of OPEs and an existing need to immediately identify a chemical flame retardant to replace the now phased out PBDEs to meet flammability regulations. In both cases of regrettable substitutions for chemical flame retardants, there was insufficient scientific evidence on the potential adverse effects of the chemical and an urgency to fill the chemical flame-retardant vacuum created by a phase out to continue to meet flammability regulations (see Figure 1). In the case of PBDEs, there was an additional oversight in the understanding that chemicals with similar structures could have similar adverse health effects if used as replacements, as well as a failure to consider their tendencies to bioaccumulate and subsequent life-cycle repercussions. For OPEs, there was an additional failure to accurately predict their exposures due to their lower half lives, resulting in an incorrect assumption that exposures to OPEs would generally be lower. An additional challenge encountered for both flame retardants was the lack of transparency from flame retardant companies on the chemicals used for 219 flame retardant mixtures. 41 This complicated early attempts to assess OPEs adverse effects on human health, with substantial time first spent characterizing flame-retardant mixtures, such as Firemaster 550. 41 Figure 1. Challenges Experienced During Prior Phase Outs Which Contributed to Use of PBDEs and OPEs As we move forward from these two instances of regrettable substitutions of chemical flame retardants and consider potential global reduction efforts of OPE exposures, an integrated plan moving forward should consider some of the previous challenges faced during chemical flame-retardant phase outs to avoid repeating similar mistakes. 41 Table 1 outlines challenges encountered during prior FR phase outs, along with recommendations compiled using a variety of sources, including recommendations from the National Academy of Sciences and existing literature examining chemical flame retardants or outlining strategies to avoid regrettable substitutions. 42 The outlined recommendations generally focus on policy and scientific research strategies to increasing transparency of chemicals being used, support creation of accessible chemical databases, and support advances in techniques and methodologies to better evaluate chemical exposures, life cycles, and prediction models. Additionally, OPE reduction efforts would benefit from identification of informed substitutions or new technologies to develop innovative flame-retardant solutions, given existing fire safety needs. There are currently a multitude of existing frameworks which OPEs PBDEs • Lack of data on potential hazards. • Focus on elimination, over functional use. • Lack of FR company transparency on chemicals. • Failure to adequately predict potential exposures. • Failure to consider life cycle concerns. • Failure to consider structural similarity. 220 integrate many of the recommendations outlined here and which could support an informed phase out of OPEs, including the National Academy of Science’s framework to guide selections of chemical alternatives; however limited research on the hazardous effects of chemicals continues to be a challenge. 23,41-43 Any future mitigation initiatives will likely require an integrated plan developed by an interdisciplinary team of experts across policy, engineering, toxicology, chemistry, biology, epidemiology, and critical stakeholders to effectively outline and execute. As epidemiologists, we can support these efforts by responding to existing calls for research within our field of expertise, including existing requests for research which links exposures and exposures mixtures to health outcomes. Table 1. Prior Phase-Out Challenges and Recommendations of Potential Solutions Moving Forward Challenges Recommendations Lack of data on potential hazards • Improved analytical methods and approaches to support research on potential hazards. * • Develop and maintain accessible chemical information database. * • Development/use of existing methodologies to monitor and identify unintended health consequences of existing chemicals. • Support research which obtains actionable data and links sources with exposures and exposure mixtures with health, especially among generally understudied environmental justice communities. * Focus on elimination over functional use • Chemical alternative assessments. • Informed substitution for essential FR uses. • Develop and research innovative FR solutions that minimize chemical use. • Create health criteria for replacement chemicals. Lack of FR transparency on chemicals used. • Increase transparency in chemical reporting requirements to include disclosures of chemical names/structures. * • Labeling system to track chemical use in products. • Increased accessibility of TSCA inventory for the scientific community. Failure to consider life cycle concerns • Interdisciplinary and systems level approaches when considering cumulative impacts of new chemicals, especially when substitutions are considered. • Encourage research investigating chemical transformation of contaminants. Failure to consider structural similarities • Improvement in Quantitative Structure Activity Relationship (QSAR) models’ ability to measure endpoints such as endocrine disruption and neurotoxicity. • Class management of chemicals rather than using a chemical-by-chemical basis. Inaccurate predictions of potential exposures • Use of integrated frameworks which examine chemical exposure pathways and possible exposure disparities into the informed substitution process. • Identification of species, materials, and partitioning phenomena which influence exposure. * • Targeted research into potential exposure pathways across a range of building types and surfaces, along with potential differential toxicity via exposure routes. * *NASEM Recommendations from Report on Indoor Air. 221 FUTURE STUDIES AND RESEARCH DIRECTIONS There are a variety of future potential investigations warranted to expand our understanding of the early impacts of OPEs on early neurodevelopment and to support future OPE mitigation efforts. For one, although we measured biomarkers of OPEs exposures using a single maternal urinary sample, future research evaluating multiple windows of exposures to OPEs in both the prenatal and postnatal period is important to developing a more thorough understanding of particularly susceptible windows in development and potential impacts of chronic OPE exposures. Additionally, our results found adverse associations between various OPE metabolites and birth outcomes, fine motor development in infancy, and neurobehavioral symptoms in childhood, but a more thorough understanding of the mechanisms underlying these associations is needed. Furthermore, evaluating exposures to OPEs in conjunction to other common indoor chemicals which may pose a risk to children’s neurodevelopment, including PBDEs, PFAS, pesticides, etc., is critical to developing a comprehensive understanding of OPEs impacts on health. Co-occurring exposures to multiple chemicals, along with early psychosocial stressors, may work synergistically to adversely impact neurodevelopmental outcomes. Thus, supervised mixtures approaches, such as BKMR, which consider a host of chemical exposures and psychosocial stressors may provide us with a more comprehensive understanding of how environmental and social stressors impact the neurodevelopmental health of both susceptible populations. This could further inform possible mixtures of environmental stressors contributing to health disparities among structurally marginalized communities who are disproportionately exposed to environmental chemical and social stressors but historically understudied. 11 Additional studies evaluating common sources of OPEs and individual level behaviors which could help minimize indoor exposures to OPEs is warranted given the current paucity of data on the topic. Prior studies have found an association between household cleaning and hand washing and reduced OPE exposures among mothers; however, the literature in this area is scare. Future interventions aimed at 222 minimizing exposures to OPEs among susceptible populations, such as pregnant women and children, would benefit from a better understanding of individual level behaviors which can decrease exposures to OPEs and common exposure sources. To date, previous research suggests that some nail polishes and sunscreens could increase exposures to TDCIPP, but more research is needed to support educational efforts and interventions to reduce exposures to OPEs. CONCLUSIONS Altogether, this dissertation found evidence of adverse effects from prenatal exposures to OPEs on early neurodevelopment, when OPEs were evaluated both independently and in mixtures, and evidence of sex-specific effects. Our results contribute evidence of the potential neurotoxic effects of OPEs used as flame retardants and plasticizers during developmentally sensitive periods of neurodevelopment. Further research is needed to examine the impacts of both prenatal and postnatal windows of OPE exposure on neurodevelopmental outcomes and to further evaluate potential mechanisms of effect and exposure sources. 223 REFERENCES 1. Monk C and Hane A. Fetal and Infant Brain–Behavior Development: Milestones & Environmental Influences. 2014. 2. Bateson P, Barker D, Clutton-Brock T, et al. Developmental plasticity and human health. Nature 2004; 430: 419-421. DOI: 10.1038/nature02725. 3. Grandjean P and Landrigan PJ. Neurobehavioural effects of developmental toxicity. The Lancet Neurology 2014; 13: 330-338. DOI: 10.1016/S1474-4422(13)70278-3. 4. Grandjean P and Landrigan PJ. Developmental neurotoxicity of industrial chemicals. Lancet 2006; 368: 2167-2178. 2006/12/19. DOI: 10.1016/s0140-6736(06)69665-7. 5. Zablotsky B, Black LI, Maenner MJ, et al. Prevalence and Trends of Developmental Disabilities among Children in the United States: 2009–2017. Pediatrics 2019; 144: e20190811. DOI: 10.1542/peds.2019-0811. 6. Blum A, Behl M, Birnbaum L, et al. Organophosphate Ester Flame Retardants: Are They a Regrettable Substitution for Polybrominated Diphenyl Ethers? Environ Sci Technol Lett 2019; 6: 638- 649. 2020/06/05. DOI: 10.1021/acs.estlett.9b00582. 7. Stapleton HM, Klosterhaus S, Eagle S, et al. Detection of organophosphate flame retardants in furniture foam and U.S. house dust. Environ Sci Technol 2009; 43: 7490-7495. 2009/10/24. DOI: 10.1021/es9014019. 8. Doherty BT, Hammel SC, Daniels JL, et al. Organophosphate Esters: Are These Flame Retardants and Plasticizers Affecting Children’s Health? Curr Environ Health Rep 2019; 6: 201-213. DOI: 10.1007/s40572-019-00258-0. 9. Yang J, Zhao Y, Li M, et al. A Review of a Class of Emerging Contaminants: The Classification, Distribution, Intensity of Consumption, Synthesis Routes, Environmental Effects and Expectation of Pollution Abatement to Organophosphate Flame Retardants (OPFRs). Int J Mol Sci 2019; 20 2019/06/20. DOI: 10.3390/ijms20122874. 10. Johnston J and Cushing L. Chemical Exposures, Health, and Environmental Justice in Communities Living on the Fenceline of Industry. Curr Environ Health Rep 2020; 7: 48-57. 2020/01/24. DOI: 10.1007/s40572-020-00263-8. 11. Buckley JP, Kuiper JR, Bennett DH, et al. Exposure to Contemporary and Emerging Chemicals in Commerce among Pregnant Women in the United States: The Environmental influences on Child Health Outcome (ECHO) Program. Environmental Science & Technology 2022; 56: 6560-6573. DOI: 10.1021/acs.est.1c08942. 12. Hernandez-Castro I, Eckel SP, Howe CG, et al. Sex-specific effects of prenatal organophosphate ester (OPE) metabolite mixtures and adverse infant birth outcomes in the maternal and developmental risks from environmental and social stressors (MADRES) pregnancy cohort. Environmental Research 2023; 226: 115703. DOI: https://doi.org/10.1016/j.envres.2023.115703. 13. Rock KD, St Armour G, Horman B, et al. Effects of Prenatal Exposure to a Mixture of Organophosphate Flame Retardants on Placental Gene Expression and Serotonergic Innervation in the Fetal Rat Brain. Toxicol Sci 2020; 176: 203-223. 2020/04/04. DOI: 10.1093/toxsci/kfaa046. 14. Hoffman K, Stapleton HM, Lorenzo A, et al. Prenatal exposure to organophosphates and associations with birthweight and gestational length. Environ Int 2018; 116: 248-254. 2018/04/27. DOI: 10.1016/j.envint.2018.04.016. 15. Luo D, Liu W, Tao Y, et al. Prenatal Exposure to Organophosphate Flame Retardants and the Risk of Low Birth Weight: A Nested Case-Control Study in China. Environ Sci Technol 2020; 54: 3375- 3385. 2020/02/29. DOI: 10.1021/acs.est.9b06026. 16. Liu W, Luo D, Xia W, et al. Prenatal exposure to halogenated, aryl, and alkyl organophosphate esters and child neurodevelopment at two years of age. J Hazard Mater 2020; 408: 124856. 2021/01/01. DOI: 10.1016/j.jhazmat.2020.124856. 224 17. Castorina R, Bradman A, Stapleton HM, et al. Current-use flame retardants: Maternal exposure and neurodevelopment in children of the CHAMACOS cohort. Chemosphere 2017; 189: 574-580. 2017/10/01. DOI: 10.1016/j.chemosphere.2017.09.037. 18. Doherty BT, Hoffman K, Keil AP, et al. Prenatal exposure to organophosphate esters and cognitive development in young children in the Pregnancy, Infection, and Nutrition Study. Environ Res 2019; 169: 33-40. 2018/11/09. DOI: 10.1016/j.envres.2018.10.033. 19. Doherty BT, Hoffman K, Keil AP, et al. Prenatal exposure to organophosphate esters and behavioral development in young children in the Pregnancy, Infection, and Nutrition Study. Neurotoxicology 2019; 73: 150-160. 2019/04/06. DOI: 10.1016/j.neuro.2019.03.007. 20. Patisaul HB, Behl M, Birnbaum LS, et al. Beyond Cholinesterase Inhibition: Developmental Neurotoxicity of Organophosphate Ester Flame Retardants and Plasticizers. Environ Health Perspect 2021; 129: 105001. 2021/10/07. DOI: 10.1289/ehp9285. 21. Morgan AB and Wilkie CA. Flame retardant polymer nanocomposites. John Wiley & Sons, 2007. 22. Vuong AM, Yolton K, Cecil KM, et al. Flame retardants and neurodevelopment: An updated review of epidemiological literature. Curr Epidemiol Rep 2020; 7: 220-236. 2021/01/08. DOI: 10.1007/s40471-020-00256-z. 23. Maertens A, Golden E and Hartung T. Avoiding Regrettable Substitutions: Green Toxicology for Sustainable Chemistry. ACS Sustainable Chemistry & Engineering 2021; 9: 7749-7758. DOI: 10.1021/acssuschemeng.0c09435. 24. Segev O, Kushmaro A and Brenner A. Environmental impact of flame retardants (persistence and biodegradability). Int J Environ Res Public Health 2009; 6: 478-491. 2009/05/15. DOI: 10.3390/ijerph6020478. 25. Visakh P and Yoshihiko A. Flame retardants: Polymer blends, composites and nanocomposites. Springer, 2015. 26. Yang F, Ding J, Huang W, et al. Particle size-specific distributions and preliminary exposure assessments of organophosphate flame retardants in office air particulate matter. Environmental science & technology 2014; 48: 63-70. 27. Hou R, Xu Y and Wang Z. Review of OPFRs in animals and humans: Absorption, bioaccumulation, metabolism, and internal exposure research. Chemosphere 2016; 153: 78-90. DOI: https://doi.org/10.1016/j.chemosphere.2016.03.003. 28. Diamond ML. Toxic chemicals as enablers and poisoners of the technosphere. The Anthropocene Review 2017; 4: 72-80. 29. Berghuis SA, Van Braeckel KN, Sauer PJ, et al. Prenatal exposure to persistent organic pollutants and cognition and motor performance in adolescence. Environ Int 2018; 121: 13-22. 30. Caspersen IH, Aase H, Biele G, et al. The influence of maternal dietary exposure to dioxins and PCBs during pregnancy on ADHD symptoms and cognitive functions in Norwegian preschool children. Environ Int 2016; 94: 649-660. 31. Ruel MVM, Bos AF, Soechitram SD, et al. Prenatal exposure to organohalogen compounds and children’s mental and motor development at 18 and 30 months of age. NeuroToxicology 2019; 72: 6-14. DOI: https://doi.org/10.1016/j.neuro.2019.01.003. 32. Rosenquist AH, Høyer BB, Julvez J, et al. Prenatal and postnatal PCB-153 and p, p′-DDE exposures and behavior scores at 5–9 years of age among children in Greenland and Ukraine. Environ Health Perspect 2017; 125: 107002. 33. Verner M-A, Hart JE, Sagiv SK, et al. Measured Prenatal and Estimated Postnatal Levels of Polychlorinated Biphenyls (PCBs) and ADHD-Related Behaviors in 8-Year-Old Children. Environ Health Perspect 2015; 123: 888-894. DOI: doi:10.1289/ehp.1408084. 34. Šovčíková E, Wimmerová S, Stremy M, et al. Simple reaction time in 8–9-year old children environmentally exposed to PCBs. Neurotoxicology 2015; 51: 138-144. 225 35. Braun JM, Yolton K, Stacy SL, et al. Prenatal environmental chemical exposures and longitudinal patterns of child neurobehavior. NeuroToxicology 2017; 62: 192-199. DOI: https://doi.org/10.1016/j.neuro.2017.07.027. 36. Zhang H, Yolton K, Webster GM, et al. Prenatal PBDE and PCB exposures and reading, cognition, and externalizing behavior in children. Environ Health Perspect 2017; 125: 746-752. 37. Eskenazi B, Fenster L, Castorina R, et al. A comparison of PBDE serum concentrations in Mexican and Mexican-American children living in California. Environ Health Perspect 2011; 119: 1442- 1448. 2011/04/15. DOI: 10.1289/ehp.1002874. 38. Stapleton HM, Klosterhaus S, Keller A, et al. Identification of flame retardants in polyurethane foam collected from baby products. Environ Sci Technol 2011; 45: 5323-5331. 2011/05/20. DOI: 10.1021/es2007462. 39. Sugeng EJ, de Cock M, Leonards PEG, et al. Association of exposure to organophosphate flame retardants and children's behavior at a median age of 18 months. Environmental Advances 2021; 5: 100077. DOI: https://doi.org/10.1016/j.envadv.2021.100077. 40. Santillo D and Johnston P. Playing with fire: the global threat presented by brominated flame retardants justifies urgent substitution. Environ Int 2003; 29: 725-734. DOI: https://doi.org/10.1016/S0160-4120(03)00115-6. 41. Howard GJ. Chemical alternatives assessment: The case of flame retardants. Chemosphere 2014; 116: 112-117. DOI: https://doi.org/10.1016/j.chemosphere.2014.02.034. 42. National Academies of Sciences E and Medicine. Why Indoor Chemistry Matters. Washington, DC: The National Academies Press, 2022, p.190. 43. Page J, Whaley P, Bellingham M, et al. A new consensus on reconciling fire safety with environmental & health impacts of chemical flame retardants. Environ Int 2023; 173: 107782. DOI: https://doi.org/10.1016/j.envint.2023.107782.

Asset Metadata

Creator Hernandez-Castro, Ixel Carolina (author)

Core Title Prenatal exposure to organophosphate esters and early neurodevelopment

Contributor Electronically uploaded by the author (provenance)

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Epidemiology

Degree Conferral Date 2023-08

Publication Date 06/08/2023

Defense Date 05/31/2023

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag BKMR,mixtures,neurodevelopment,OAI-PMH Harvest,OPEs,OPFRs,organophosphate esters

Format theses (aat)

Language English

Advisor Bastain, Theresa M. (committee chair), Aung, Max T. (committee member), Breton, Carrie V. (committee member), Eckel, Sandrah P. (committee member), Grubbs, Brendan (committee member)

Creator Email ixelhern@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC113169914

Unique identifier UC113169914

Identifier etd-HernandezC-11942.pdf (filename)

Legacy Identifier etd-HernandezC-11942

Document Type Dissertation

Format theses (aat)

Rights Hernandez-Castro, Ixel Carolina

Internet Media Type application/pdf

Type texts

Source 20230612-usctheses-batch-1054 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email uscdl@usc.edu

Abstract (if available)

Abstract Early neurodevelopment begins in-utero with the formation of integral neurological structures and mechanisms and plays a critical role in supporting children’s lifelong health. According to previous studies, ubiquitous environmental chemicals, such as organophosphate esters (OPEs) used as flame retardants and plasticizers in a variety of consumer products, may pose a neurotoxic risk at environmentally relevant doses. However, there is limited observational evidence evaluating the association between prenatal exposures to OPEs and early neurodevelopment, particularly among populations historically underrepresented in the biomedical sciences. Additionally, there is scarce evidence evaluating the impacts of co-occurring OPEs on early neurodevelopment, along with many OPEs whose impacts on neurodevelopment are not well understood. We investigated the impacts of prenatal OPE exposures on early neurodevelopmental outcomes among a predominately low-income and Hispanic pregnancy cohort of participants residing in Los Angeles, California, the Maternal and Developmental Risks from Environmental and Social Stressors (MADRES) cohort. We evaluated the association between nine OPE metabolites measured in urine samples collected during a third trimester visit and outcomes associated with neurodevelopmental health, including gestational age (GA) at birth, birthweight, motor development in infancy, and neurobehavioral symptoms in early childhood. Altogether, we found evidence of potentially neurotoxic impacts from prenatal OPE exposures on development, with various sex specific associations, and interesting mixtures associations, which highlight the value of evaluating co-occurring OPE exposures on children’s health outcomes. These findings can potentially inform future reduction efforts to minimize adverse effects of prenatal OPEs on neurodevelopmental health but would benefit from further research on individual sources of OPEs along with more research using multiple measurements of OPEs across prenatal and postnatal windows.